NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16

\n\n $\"Chat\"$ \n\n\n $\"Paper\"$ \n\n\n $\"Pre-Training$ \n\n\n $\"Post-Training$ \n\n

\n \n $\"Homepage\"$ \n \n\n $\"Discord\"$ \n \n

\n \n $\"License\"$ \n \n

$\"\"$

Model Summary

\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n

tr:nth-child(odd)]:bg-surface-sunken\">


Total Parameters	550B (55B active)
Architecture	LatentMoE - Mamba-2 + MoE + Attention hybrid with Multi-Token Prediction (MTP)
Context Length	Up to 1M tokens
Minimum GPU Requirement	8x GB200/B200/GB300/B300, 16x H100, 8x H200
Supported Languages	English, French, Spanish, Italian, German, Japanese, Korean, Hindi, Korean, Brazilian Portuguese, and Chinese
Best For	Frontier reasoning, complex agentic workflows, long-context analysis, tool use, multilingual reasoning, high-stakes RAG
Reasoning Mode	Configurable on/off via chat template (`enable_thinking=True/False`)
License	OpenMDW License Agreement, version 1.1
Release Date	June 4, 2026

Quick Start

For more details on how to deploy and use the model - see the Quick Start Guide below!

\n
For running Nemotron 3 Ultra on a smaller footprint, please see: NVIDIA-Nemotron-3-Ultra-550B-A55B-NVFP4
\n

Model Overview

Model Developer: NVIDIA Corporation

Model Dates: December 2025 - April 2026

Data Freshness:

The post-training data has a cutoff date of May 2026.
The pre-training data has a cutoff date of September 2025.

What is Nemotron?

NVIDIA Nemotron™ is a family of open models with open weights, training data, and recipes, delivering leading efficiency and accuracy for building specialized AI agents.

Description

Nemotron-3-Ultra-550B-A55B-BF16 is a frontier-scale large language model (LLM) trained by NVIDIA, designed to deliver strong agentic, reasoning, and conversational capabilities. It is optimized for the most demanding workloads, including complex multi-step agents, long-context analysis, and high-accuracy reasoning over code, math, and science. Like other models in the family, it responds to user queries and tasks by first generating a reasoning trace and then concluding with a final response. The model's reasoning capabilities can be configured through a flag in the chat template.

The model employs a hybrid Latent Mixture-of-Experts (LatentMoE) architecture, utilizing interleaved Mamba-2 and MoE layers, along with select Attention layers. Like the Super model, the Ultra model incorporates Multi-Token Prediction (MTP) layers for faster text generation and improved quality, and it is trained using an NVFP4 pre-training recipe to maximize compute efficiency. The model has 55B active parameters and 550B parameters in total.

The supported languages include: English, French, Spanish, Italian, German, Japanese, Korean, Hindi, Korean, Brazilian Portuguese, and Chinese.

This model is ready for commercial and non-commercial use.

License/Terms of Use

Governing Download Terms: Use of this model is governed by the OpenMDW License Agreement, version 1.1 (OpenMDW-1.1).

Benchmarks

tr:nth-child(odd)]:bg-surface-sunken\">

Benchmark	N-3-Ultra 550B-A55B	MiniMax-2.7 230B-A10B	GLM-5.1 744B-A40B	Kimi-K2.6 1T-A32B	Qwen-3.5 397B-17B	DS-v4-Pro 1.6T-A49B	DS-v4-Flash 284B-A13B
Agentic
Terminal Bench 2.1	56.4	55.5	59.3	67.2	49.9	49.2	54.2
GDPVal	46.7	47.6	54.7	50.4	34.6	54.6	50.2
SWE-Bench Verified	71.9	72.2	73.8	69.5	69.9	74.0	72.4
SWE-Bench Multilingual	67.7	69.2	73.8	65.9	67.7	71.9	72.1
ProfBench (Search)	56.0	52.0	46.0	56.0	53.0	59.9	57.0
PinchBench	90.0	77.6	81.2	90.2	86.6	88.6	91.3
TauBench V3
Airline	81.5	75.3	85.0	85.8	76.5	80.8	80.8
Retail	86.4	84.9	84.1	82.9	88.5	88.9	89.1
Telecom	92.9	89.6	96.9	97.8	98.0	96.3	98.3
Banking	22.6	14.6	12.8	23.1	20.9	25.9	26.7
Average	70.9	66.1	69.7	72.4	71.0	73.2	73.7
BrowseComp	44.4	54.1	59.4	61.3	40.5	59.4	46.9
Vals.ai Financial Agent 1.1
without web search	60.1	51.3	60.2	54.0	61.3	58.9	58.4
with web search	53.7	50.5	60.7	58.8	59.0	62.3	60.1
Reasoning and Knowledge
IOI 2025	570.0	--	456.5	585.0	441.3	580.1	--
LiveCodeBench (v6)	89.0	77.2	85.7	90.2	79.3	92.5	90.9
IMOAnswerBench (no tools)	88.6	68.3	86.8	91.1	83.1	93.0	91.1
IMOAnswerBench (with tools)	92.3	75.1	91.1	93.71	84.51	85.4	89.6
Apex-Shortlist (no tools)	74.9	28.9	71.1	77.4	61.4	85.8	82.4
Apex-Shortlist (with tools)	84.8	51.9	79.0	73.2	60.4	86.5	82.0
GPQA (no tools)	87.0	86.6	86.1	91.0	87.1	87.8	88.5
SciCode (subtask)	44.6	38.3	47.7	52.0	48.0	50.5	48.2
HLE (no tools)	26.7	23.1	27.2	34.8	28.5	37.7	32.2
HLE (with tools)	37.4	--	50.4	54.0	48.3	48.2	45.1
CritPt (no tools)	3.1	0.6	3.7	9.1	2.4	14.0	10.6
MMLU-Pro	86.8	81.9	85.9	88.1	88.3	87.5	86.4
OmniScience Accuracy	24.1	20.5	31.3	35.5	35.9	46.8	39.9
OmniScience Non-Hallucination	78.7	74.4	66.8	67.1	7.4	5.7	2.8
Chat & Instruction Following
IFBench (prompt loose)	81.7	74.6	76.6	73.7	78.2	79.1	82.0
Multi-Challenge	63.8	42.5	63.0	63.1	63.9	64.1	63.5
Long Context
AA-LCR	65.4	69.8	66.9	70.2	68.3	67.3	62.7
RULER (1M)	94.7	--	--	--	90.1	94.2	87.7
Longbench v2 (≤ 1M)	61.9	--	--	--	68.9	62.1	57.0
Multilingual
MMLU-ProX (avg en/de/fr/es/it/ja/zh/hi/pt/ko)	83.0	78.4	85.8	85.0	86.4	85.6	84.3
WMT24++ (en→xx)	83.7	82.8	84.4	84.5	86.8	85.9	85.9

All evaluation results were collected via Nemo Evaluator SDK. We used three main evaluation harnesses: Nemo Gym, Nemo Skills, and Harbor with extended sandboxing support via AWS ECS on Nemo Evaluator. In addition, the evaluations also used dedicated open-source packaged containers for ScaleAI Multi Challenge Multi Turn Instruction Following and KernelBench. For reproducibility purposes, more details on the evaluation settings and pinned containers can be found in the Nemo Evaluator SDK examples folder and the reproducibility tutorial for Nemotron 3 Ultra.

The following benchmarks are not onboarded yet in our open source tools and for these we used either their official open source implementation or otherwise an internal scaffolding that we plan to open source in the future: BrowseComp with Search, Tau Bench 3, ProfBench with Search, PinchBench, Vals.ai, LongBench v2.

Deployment Geography: Global

Use Case

NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16 is a frontier-scale general purpose reasoning and chat model intended to be used in English, Code, and supported multilingual contexts. This model is optimized for complex agentic workflows, long-context reasoning, and high-stakes analytical workloads. It is intended to be used by developers designing AI Agent systems, chatbots, RAG systems, and other AI-powered applications. This model is also suitable for complex instruction-following tasks and long-context reasoning over very large documents and codebases.

Release Date

Hugging Face - 06/04/2026 via Hugging Face

Reference(s)

Model Architecture

Architecture Type: Mamba2-Transformer Hybrid Latent Mixture of Experts (LatentMoE) with Multi-Token Prediction (MTP)
Network Architecture: Nemotron Hybrid LatentMoE
Number of model parameters: 550B Total / 55B Active

Model Design

The model utilizes the LatentMoE architecture, where tokens are projected into a smaller latent dimension for expert routing and computation, improving accuracy per byte. The Ultra model is pre-trained using an NVFP4 recipe — sharing the quantization-aware pre-training approach pioneered in the Nemotron 3 family. The majority of linear layers use NVFP4 for weights, activations, and gradients, while select layers (including latent projections, MTP layers, QKV/attention projections, and embeddings) are maintained in BF16 or MXFP8 for training stability. The model includes Multi-Token Prediction (MTP) layers using a shared-weight design across prediction heads. This improves training signal quality, enables faster inference via native speculative decoding, and supports more stable autoregressive drafting at longer draft lengths compared to independently trained offset heads.

Training Methodology

Stage 1: Pre-Training

NVIDIA-Nemotron-3-Ultra-550B-A55B-Base-BF16 model was pre-trained for approximately 20T tokens using crawled and synthetic code, math, science, and general knowledge data. Training leveraged an NVFP4 recipe for efficiency. All datasets are disclosed in the Training and Evaluation Datasets section of this document. Major portions of the pre-training corpus are released in the Nemotron-Pre-Training-Datasets collection.
Software used for pre-training: Megatron-LM

Stage 2: Supervised Fine-Tuning

The model was further fine-tuned on synthetic code, math, science, tool calling, instruction following, structured outputs, and general knowledge data. This stage incorporated data designed to support long-range retrieval and multi-document aggregation. All datasets are disclosed in the Training and Evaluation Datasets section of this document. Major portions of the fine-tuning corpus are released in the Nemotron-Post-Training-v3 collection. Data Designer is one of the libraries used to prepare these corpora.

Stage 3: Reinforcement Learning

The model underwent multi-environment reinforcement learning using asynchronous GRPO (Group Relative Policy Optimization) across math, code, science, instruction following, multi-step tool use, multi-turn conversations, and structured output environments. It utilized an asynchronous RL architecture that fully decouples training from inference across separate GPU devices, leveraging in-flight weight updates and MTP to accelerate rollout generation. Conversational quality was further refined through RLHF. All datasets are disclosed in the Training and Evaluation Datasets section of this document. The RL environments and datasets are released as part of NeMo Gym.
Software used for reinforcement learning: NeMo RL, NeMo Gym

Stage 4: Multi-Domain On-Policy Distillation (MOPD)

The model underwent Multi-Domain On-Policy Distillation (MOPD) to improve reasoning across many task types while staying efficient. This technique uses strong teacher models to guide training on the model's own generated attempts (on-policy rollouts), helping recover accuracy and improve performance across coding, math, instruction following, tool use, and agentic workflows. By distilling teacher signal onto the student's own trajectories rather than offline traces, MOPD better aligns the student's behavior with what it would actually produce at inference time, yielding stronger gains than purely off-policy distillation.

NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16 model is a result of the above work.

The end-to-end training recipe is available in the NVIDIA Nemotron Developer Repository. Evaluation results can be replicated using the NeMo Evaluator SDK. Data Designer is one of the libraries used to prepare the pre and post training datasets. More details on the datasets and synthetic data generation methods can be found in the technical report NVIDIA Nemotron 3 Ultra Technical Report.

Input

Input Type(s): Text
Input Format(s): String
Input Parameters: One-Dimensional (1D): Sequences
Other Properties Related to Input: Maximum context length up to 1M tokens. Supported languages include: English, French, Spanish, Italian, German, Japanese, Korean, Hindi, Korean, Brazilian Portuguese, and Chinese

Output

Output Type(s): Text
Output Format: String
Output Parameters: One-Dimensional (1D): Sequences
Other Properties Related to Output: Maximum context length up to 1M tokens

Our AI models are designed and optimized to run on NVIDIA GPU-accelerated systems. By leveraging NVIDIA's hardware (e.g. GPU cores) and software frameworks (e.g., CUDA libraries), the model achieves faster training and inference times compared to CPU-only solutions.

Software Integration

Runtime Engine(s): NeMo 26.04.01
Supported Hardware Microarchitecture Compatibility: NVIDIA Ampere - A100; NVIDIA Blackwell; NVIDIA Hopper - H100-80GB
Operating System(s): Linux

The integration of foundation and fine-tuned models into AI systems requires additional testing using use-case-specific data to ensure safe and effective deployment. Following the V-model methodology, iterative testing and validation at both unit and system levels are essential to mitigate risks, meet technical and functional requirements, and ensure compliance with safety and ethical standards before deployment.

Model Version(s)

v1.0 - GA

The Ultra BF16 checkpoint is a frontier-scale model. The minimum recommended hardware is:

Single-node: 8× B200 (≈1.5 TB aggregate HBM — fits BF16 weights plus KV cache with headroom)
Multi-node: ≥8 GPUs across H100 / H200 / GB200 / GB300, orchestrated with Ray v2

All deployment snippets below default to port 8000, with chunked prefill and MTP (5 speculative tokens) enabled.

Multi-Node Setup with Ray (recommended for multi-node BF16)

The recommended multi-processing backend for multi-node BF16 deployments is Ray v2. Below is a template for launching a Ray cluster:

\nexport RAY_PORT=6379\nexport RAY_ADDRESS=${RAY_HEAD_IP}:${RAY_PORT}\n\n# Start Ray head node (vLLM/SGLang will run on this node)\nray start --head --node-ip-address=${RAY_HEAD_IP} --port=${RAY_PORT}\n\n# Start Ray worker node(s)\nray start --address=${RAY_HEAD_IP}:${RAY_PORT} --block\n\n# Verify Ray cluster is ready\nray status --address=${RAY_HEAD_IP}:${RAY_PORT}\n\">

# Set the IP for the head node in RAY_HEAD_IP\nexport RAY_HEAD_IP=<head_node_ip>\nexport RAY_PORT=6379\nexport RAY_ADDRESS=${RAY_HEAD_IP}:${RAY_PORT}\n\n# Start Ray head node (vLLM/SGLang will run on this node)\nray start --head --node-ip-address=${RAY_HEAD_IP} --port=${RAY_PORT}\n\n# Start Ray worker node(s)\nray start --address=${RAY_HEAD_IP}:${RAY_PORT} --block\n\n# Verify Ray cluster is ready\nray status --address=${RAY_HEAD_IP}:${RAY_PORT}\n

ray[cgraph] is required: uv pip install \"ray[cgraph]\"

vLLM

Recommended container: vllm/vllm-openai:v0.22.0.

For more detailed information, please see this cookbook.

export MODEL_CKPT=PATH/TO/MODEL/CHECKPOINT\n

8× B200 single-node deployment:

docker run -d --name nemotron-ultra-vllm \\\n  --gpus all \\\n  --ipc=host \\\n  --network=host \\\n  --shm-size=16g \\\n  --ulimit memlock=-1 \\\n  --ulimit stack=67108864 \\\n  -v $MODEL_CKPT:/model:ro \\\n  -e VLLM_WORKER_MULTIPROC_METHOD=spawn \\\n  -e SAFETENSORS_FAST_GPU=1 \\\n  -e NVIDIA_TF32_OVERRIDE=1 \\\n  -e VLLM_LOGGING_LEVEL=INFO \\\n  vllm/vllm-openai:v0.22.0 \\\n  /model \\\n  --host 0.0.0.0 \\\n  --port 8000 \\\n  --served-model-name nvidia/nemotron-3-ultra \\\n  --trust-remote-code \\\n  --tensor-parallel-size 8 \\\n  --enable-expert-parallel \\\n  --dtype bfloat16 \\\n  --max-model-len 262144 \\\n  --gpu-memory-utilization 0.90 \\\n  --max-num-seqs 16 \\\n  --max-num-batched-tokens 32768 \\\n  --enable-chunked-prefill \\\n  --enable-prefix-caching \\\n  --reasoning-parser nemotron_3 \\\n  --enable-auto-tool-choice \\\n  --tool-call-parser qwen3_coder \\\n  --mamba-ssm-cache-dtype float16 \\   \n  --mamba-backend flashinfer \\     \n  --enable-mamba-cache-stochastic-rounding \\   \n  --mamba-cache-philox-rounds 5 \\\n  --speculative-config '{\"method\": \"nemotron_h_mtp\", \"num_speculative_tokens\": 5}' \\\n  --model-loader-extra-config '{\"enable_multithread_load\": true, \"num_threads\": 96}'\n

Multi-node deployment (e.g. 2× 4×GB300 with Ray):

After launching the Ray head and worker per the multi-node setup above:

# Run on Ray head node\nvllm serve $MODEL_CKPT \\\n  --host 0.0.0.0 \\\n  --port 8000 \\\n  --served-model-name nvidia/nemotron-3-ultra \\\n  --tensor-parallel-size 8 \\\n  --distributed-executor-backend ray \\\n  --trust-remote-code \\\n  --dtype bfloat16 \\\n  --gpu-memory-utilization 0.90 \\\n  --max-model-len 262144 \\\n  --max-num-seqs 256 \\\n  --max-num-batched-tokens 32768 \\\n  --enable-chunked-prefill \\\n  --enable-prefix-caching \\\n  --reasoning-parser nemotron_3 \\\n  --mamba-ssm-cache-dtype float16 \\   \n  --mamba-backend flashinfer \\     \n  --enable-mamba-cache-stochastic-rounding \\   \n  --mamba-cache-philox-rounds 5 \\\n  --enable-auto-tool-choice \\\n  --tool-call-parser qwen3_coder \\\n  --speculative-config '{\"method\": \"nemotron_h_mtp\", \"num_speculative_tokens\": 5}' \\\n  --kv-cache-dtype fp8 \\\n  --model-loader-extra-config '{\"enable_multithread_load\": true, \"num_threads\": 96}' \\\n  --compilation-config '{\"pass_config\": {\"fuse_allreduce_rms\": false}}' \\\n  --distributed-timeout-seconds 3600\n

Context length defaults to 256k above. To use up to 1M, set VLLM_ALLOW_LONG_MAX_MODEL_LEN=1 and --max-model-len 1048576.

Useful environment variables: VLLM_FLASHINFER_ALLREDUCE_BACKEND=trtllm, VLLM_FLASHINFER_MOE_BACKEND=latency (TRTLLM-Gen) or VLLM_FLASHINFER_MOE_BACKEND=throughput (CUTLASS).

SGLang

Container (tested on 8× B200):

docker pull lmsysorg/sglang:v0.5.11\n

For more detailed information, please see this cookbook.

8× B200 single-node deployment (BF16, chunked prefill + MTP on by default):

docker run -d --name nemotron-ultra-sglang \\\n  --gpus all \\\n  --cap-add SYS_NICE \\\n  --ipc=host \\\n  --network=host \\\n  --shm-size=16g \\\n  --ulimit memlock=-1 \\\n  --ulimit stack=67108864 \\\n  -v $MODEL_CKPT:/model:ro \\\n  -e SAFETENSORS_FAST_GPU=1 \\\n  -e NVIDIA_TF32_OVERRIDE=1 \\\n  -e SGLANG_DISABLE_DEEP_GEMM=1 \\\n  lmsysorg/sglang:v0.5.11 \\\n  python3 -m sglang.launch_server \\\n  --model-path /model \\\n  --host 0.0.0.0 \\\n  --port 8000 \\\n  --served-model-name nvidia/nemotron-3-ultra \\\n  --tp-size 8 \\\n  --ep-size 8 \\\n  --context-length 262144 \\\n  --mem-fraction-static 0.85 \\\n  --chunked-prefill-size 32768 \\\n  --fp8-gemm-backend triton \\\n  --moe-runner-backend triton \\\n  --mamba-scheduler-strategy no_buffer \\\n  --disable-piecewise-cuda-graph \\\n  --reasoning-parser nemotron_3 \\\n  --tool-call-parser qwen3_coder \\\n  --speculative-algorithm EAGLE \\\n  --speculative-num-steps 5 \\\n  --speculative-eagle-topk 1 \\\n  --speculative-num-draft-tokens 5 \\\n  --trust-remote-code \\\n  --log-level info\n

Context length defaults to 256k above. To use up to 1M, set SGLANG_ALLOW_OVERWRITE_LONGER_CONTEXT_LEN=1 and --context-length 1048576.

Tool calls + reasoning parsing: when calling the chat completions endpoint with tools, you must set \"chat_template_kwargs\": {\"enable_thinking\": true, \"force_nonempty_content\": true} in the request body to parse both reasoning and tool calls correctly.

TensorRT-LLM

Important Note: Current support for Nemotron 3 Ultra is limited to NVIDIA Blackwell architecture (including B200/B300 and GB200/GB300). While support for NVIDIA Hopper systems is planned, it is not currently available.

Container:

docker pull nvcr.io/nvidia/tensorrt-llm/release:1.3.0rc17\n

For more detailed information, please see this cookbook.

8xB200:

./extra-llm-api-config.yml << EOF\nbackend: pytorch\ntrust_remote_code: true\ntensor_parallel_size: 8\npipeline_parallel_size: 1\ncontext_parallel_size: 1\ngpus_per_node: 8\nmoe_expert_parallel_size: 1\ndisable_overlap_scheduler: false\n\ncuda_graph_config:\n enable_padding: true\n max_batch_size: 256\n\nenable_chunked_prefill: true\nenable_attention_dp: false\nmax_batch_size: 256\nmax_seq_len: null\nmax_num_tokens: 32768\nnum_postprocess_workers: 4\n\nkv_cache_config:\n enable_block_reuse: false\n max_tokens: null\n max_attention_window: null\n sink_token_length: null\n free_gpu_memory_fraction: 0.75\n host_cache_size: null\n cross_kv_cache_fraction: null\n secondary_offload_min_priority: null\n event_buffer_max_size: 0\n attention_dp_events_gather_period_ms: 5\n enable_partial_reuse: true\n copy_on_partial_reuse: true\n use_uvm: false\n max_gpu_total_bytes: 0\n iteration_stats_interval: 1\n dtype: fp8\n tokens_per_block: 32\n mamba_state_cache_interval: 256\n use_kv_cache_manager_v2: false\n max_util_for_resume: 0.95\n\nmoe_config:\n backend: TRTLLM\n max_num_tokens: null\n load_balancer: null\n disable_finalize_fusion: false\n use_low_precision_moe_combine: false\nEOF\n\n\nTLLM_ALLOW_LONG_MAX_MODEL_LEN=1 trtllm-serve \\\n \\\n--max_batch_size 256 \\\n--tp_size 8 --ep_size 1 \\\n--max_num_tokens 32768 \\\n--trust_remote_code \\\n--reasoning_parser nano-v3 \\\n--tool_parser qwen3_coder \\\n--chat_template $MODEL_DIR/chat_template.jinja \\\n--extra_llm_api_options extra-llm-api-config.yml\n\">

cat > ./extra-llm-api-config.yml << EOF\nbackend: pytorch\ntrust_remote_code: true\ntensor_parallel_size: 8\npipeline_parallel_size: 1\ncontext_parallel_size: 1\ngpus_per_node: 8\nmoe_expert_parallel_size: 1\ndisable_overlap_scheduler: false\n\ncuda_graph_config:\n  enable_padding: true\n  max_batch_size: 256\n\nenable_chunked_prefill: true\nenable_attention_dp: false\nmax_batch_size: 256\nmax_seq_len: null\nmax_num_tokens: 32768\nnum_postprocess_workers: 4\n\nkv_cache_config:\n  enable_block_reuse: false\n  max_tokens: null\n  max_attention_window: null\n  sink_token_length: null\n  free_gpu_memory_fraction: 0.75\n  host_cache_size: null\n  cross_kv_cache_fraction: null\n  secondary_offload_min_priority: null\n  event_buffer_max_size: 0\n  attention_dp_events_gather_period_ms: 5\n  enable_partial_reuse: true\n  copy_on_partial_reuse: true\n  use_uvm: false\n  max_gpu_total_bytes: 0\n  iteration_stats_interval: 1\n  dtype: fp8\n  tokens_per_block: 32\n  mamba_state_cache_interval: 256\n  use_kv_cache_manager_v2: false\n  max_util_for_resume: 0.95\n\nmoe_config:\n  backend: TRTLLM\n  max_num_tokens: null\n  load_balancer: null\n  disable_finalize_fusion: false\n  use_low_precision_moe_combine: false\nEOF\n\n\nTLLM_ALLOW_LONG_MAX_MODEL_LEN=1 trtllm-serve  \\\n<bf16_ckpt> \\\n--max_batch_size 256 \\\n--tp_size 8 --ep_size 1 \\\n--max_num_tokens 32768 \\\n--trust_remote_code \\\n--reasoning_parser nano-v3 \\\n--tool_parser qwen3_coder \\\n--chat_template $MODEL_DIR/chat_template.jinja \\\n--extra_llm_api_options extra-llm-api-config.yml\n

Long-context configuration:

For long-context benchmarking, set TLLM_ALLOW_LONG_MAX_MODEL_LEN=1 as an environment variable and add --max_seq_len <seq_len> as the desired maximum context length. The MTP speculative_config block above carries over unchanged — on rc16, max_draft_len is the authoritative field and num_nextn_predict_layers is treated as deprecated.

API Client

The examples below use the OpenAI-compatible client and work with any of the serving backends above.

NOTE: For coding agents add the following to the API call - extra_body={\"chat_template_kwargs\": {\"force_nonempty_content\": True}}

from openai import OpenAI\nclient = OpenAI(base_url=\"http://localhost:8000/v1\", api_key=\"EMPTY\")\nMODEL = \"nvidia/nemotron-3-ultra\"\n

Reasoning ON (default)

response = client.chat.completions.create(\n    model=MODEL,\n    messages=[{\"role\": \"user\", \"content\": \"Write a haiku about GPUs\"}],\n    max_tokens=16000,\n    temperature=1.0,\n    top_p=0.95,\n    extra_body={\"chat_template_kwargs\": {\"enable_thinking\": True}}\n)\nprint(response.choices[0].message.content)\n

Reasoning OFF

response = client.chat.completions.create(\n    model=MODEL,\n    messages=[{\"role\": \"user\", \"content\": \"What is the capital of Japan?\"}],\n    max_tokens=16000,\n    temperature=1.0,\n    top_p=0.95,\n    extra_body={\"chat_template_kwargs\": {\"enable_thinking\": False}}\n)\nprint(response.choices[0].message.content)\n

Medium-effort reasoning

Uses significantly fewer reasoning tokens than full thinking mode. Recommended as a starting point before tuning explicit token budgets.

response = client.chat.completions.create(\n    model=MODEL,\n    messages=[{\"role\": \"user\", \"content\": \"What is the capital of Japan?\"}],\n    max_tokens=16000,\n    temperature=1.0,\n    top_p=0.95,\n    extra_body={\"chat_template_kwargs\": {\"enable_thinking\": True, \"medium_effort\": True}}\n)\nprint(response.choices[0].message.content)\n

Tool calling with reasoning (SGLang requires explicit chat template kwargs)

response = client.chat.completions.create(\n    model=MODEL,\n    messages=[{\"role\": \"user\", \"content\": \"What's the weather in New York?\"}],\n    tools=[{\n        \"type\": \"function\",\n        \"function\": {\n            \"name\": \"get_weather\",\n            \"description\": \"Get the current weather for a city.\",\n            \"parameters\": {\n                \"type\": \"object\",\n                \"properties\": {\n                    \"city\": {\"type\": \"string\"},\n                    \"unit\": {\"type\": \"string\", \"enum\": [\"celsius\", \"fahrenheit\"]}\n                },\n                \"required\": [\"city\"]\n            }\n        }\n    }],\n    tool_choice=\"required\",\n    max_tokens=256,\n    temperature=1.0,\n    top_p=0.95,\n    extra_body={\"chat_template_kwargs\": {\"enable_thinking\": True, \"force_nonempty_content\": True}}\n)\n

OpenCode

OpenCode is an AI coding agent that runs in your terminal. It connects to any OpenAI-compatible endpoint, making it compatible with all three serving backends above (vLLM, SGLang, and TensorRT-LLM).

Create or update your ~/.config/opencode/opencode.json:

{\n    \"$schema\": \"https://opencode.ai/config.json\",\n    \"model\": \"local/nvidia-nemotron-3-ultra\",\n    \"provider\": {\n        \"local\": {\n            \"npm\": \"@ai-sdk/openai-compatible\",\n            \"name\": \"local_backend\",\n            \"options\": {\n                \"baseURL\": \"http://localhost:8000/v1\",\n                \"apiKey\": \"EMPTY\"\n            },\n            \"models\": {\n                \"nvidia-nemotron-3-ultra\": {\n                    \"name\": \"nvidia/nemotron-3-ultra\",\n                    \"limit\": {\n                        \"context\": 1000000,\n                        \"output\": 32768\n                    }\n                }\n            }\n        }\n    },\n    \"agent\": {\n        \"build\": {\n            \"temperature\": 1.0,\n            \"top_p\": 0.95,\n            \"max_tokens\": 32000\n        },\n        \"plan\": {\n            \"temperature\": 1.0,\n            \"top_p\": 0.95,\n            \"max_tokens\": 32000\n        }\n    }\n}\n

All backends above default to port 8000, so the baseURL works as-is for vLLM, SGLang, and TensorRT-LLM.

To learn more about other supported agent scaffolds - check out this resource

Advanced: Budget-Controlled Reasoning

Set a hard token ceiling on the reasoning trace using reasoning_budget. The model will attempt to close the trace at the next newline before the budget is hit; if none is found within 500 tokens it closes abruptly at reasoning_budget + 500.

Dict[str, Any]:\n assert max_tokens > reasoning_budget, (\n f"reasoning_budget must be less than max_tokens. "\n f"Got {max_tokens=} and {reasoning_budget=}"\n )\n\n # Step 1: generate the reasoning trace up to the budget\n response = self.client.chat.completions.create(\n model=model, messages=messages, max_tokens=reasoning_budget, **kwargs\n )\n reasoning_content = response.choices[0].message.content\n if "" not in reasoning_content:\n reasoning_content = f"{reasoning_content}.\\n\\n\\n"\n\n reasoning_tokens_len = len(\n self.tokenizer.encode(reasoning_content, add_special_tokens=False)\n )\n remaining_tokens = max_tokens - reasoning_tokens_len\n assert remaining_tokens > 0, (\n f"No tokens remaining for response ({remaining_tokens=}). "\n "Increase max_tokens or lower reasoning_budget."\n )\n\n # Step 2: continue from the closed reasoning trace\n messages.append({"role": "assistant", "content": reasoning_content})\n prompt = self.tokenizer.apply_chat_template(\n messages, tokenize=False, continue_final_message=True\n )\n response = self.client.completions.create(\n model=model, prompt=prompt, max_tokens=remaining_tokens, **kwargs\n )\n\n return {\n "reasoning_content": reasoning_content.strip().strip("").strip(),\n "content": response.choices[0].text,\n "finish_reason": response.choices[0].finish_reason,\n }\n\">

from typing import Any, Dict, List\nimport openai\nfrom transformers import AutoTokenizer\n\n\nclass ThinkingBudgetClient:\n    def __init__(self, base_url: str, api_key: str, tokenizer_name_or_path: str):\n        self.tokenizer = AutoTokenizer.from_pretrained(tokenizer_name_or_path)\n        self.client = openai.OpenAI(base_url=base_url, api_key=api_key)\n\n    def chat_completion(\n        self,\n        model: str,\n        messages: List[Dict[str, Any]],\n        reasoning_budget: int = 512,\n        max_tokens: int = 1024,\n        **kwargs,\n    ) -> Dict[str, Any]:\n        assert max_tokens > reasoning_budget, (\n            f\"reasoning_budget must be less than max_tokens. \"\n            f\"Got {max_tokens=} and {reasoning_budget=}\"\n        )\n\n        # Step 1: generate the reasoning trace up to the budget\n        response = self.client.chat.completions.create(\n            model=model, messages=messages, max_tokens=reasoning_budget, **kwargs\n        )\n        reasoning_content = response.choices[0].message.content\n        if \"</think>\" not in reasoning_content:\n            reasoning_content = f\"{reasoning_content}.\\n\\n</think>\\n\"\n\n        reasoning_tokens_len = len(\n            self.tokenizer.encode(reasoning_content, add_special_tokens=False)\n        )\n        remaining_tokens = max_tokens - reasoning_tokens_len\n        assert remaining_tokens > 0, (\n            f\"No tokens remaining for response ({remaining_tokens=}). \"\n            \"Increase max_tokens or lower reasoning_budget.\"\n        )\n\n        # Step 2: continue from the closed reasoning trace\n        messages.append({\"role\": \"assistant\", \"content\": reasoning_content})\n        prompt = self.tokenizer.apply_chat_template(\n            messages, tokenize=False, continue_final_message=True\n        )\n        response = self.client.completions.create(\n            model=model, prompt=prompt, max_tokens=remaining_tokens, **kwargs\n        )\n\n        return {\n            \"reasoning_content\": reasoning_content.strip().strip(\"</think>\").strip(),\n            \"content\": response.choices[0].text,\n            \"finish_reason\": response.choices[0].finish_reason,\n        }\n

Example usage (32-token reasoning budget):

client = ThinkingBudgetClient(\n    base_url=\"http://localhost:8000/v1\",\n    api_key=\"EMPTY\",\n    tokenizer_name_or_path=\"nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16\",\n)\n\nresult = client.chat_completion(\n    model=\"nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16\",\n    messages=[\n        {\"role\": \"system\", \"content\": \"You are a helpful assistant. /think\"},\n        {\"role\": \"user\", \"content\": \"What is 2+2?\"},\n    ],\n    reasoning_budget=32,\n    max_tokens=512,\n    temperature=1.0,\n    top_p=0.95,\n)\nprint(result)\n

Training and Evaluation Datasets

Training

Data Modality: Text
\nThe total size: 53.8 TiB (14.8 trillion tokens)
\nTotal number of datasets: 226
\nDataset partition: Training [100%], testing [0%], validation [0%]
\nTime period for training data collection: 2013 to 2026
\nTime period for testing data collection: 2013 to 2026
\nTime period for validation data collection: 2013 to 2026
\nData Collection Method by dataset: Hybrid: Automated, Human, Synthetic
\nLabeling Method by dataset: Hybrid: Automated, Human, Synthetic

NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16 is pre-trained on a large corpus of high-quality curated and synthetically-generated data. It is trained in the English language, as well as 11 other languages and 43 programming languages. Our sources cover a variety of document types such as: webpages, dialogue, articles, and other written materials. The corpus spans domains including legal, math, science, finance, and more. We also include a small portion of question-answering, and alignment style data to improve model accuracy. The model was pre-trained for approximately 20 trillion tokens.

The post-training corpus for NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16 consists of high-quality curated and synthetically-generated data. Primary languages used for post-training include English, French, Spanish, Italian, German, Japanese, Korean, Hindi, Korean, Brazilian Portuguese, and Chinese.

These datasets, such as FinePDFs, EssentialWeb, HotpotQA, SQuAD, and HelpSteer3, do not collectively or exhaustively represent all demographic groups (and proportionally therein). For instance, these datasets do not contain explicit mentions of demographic classes such as age, gender, or ethnicity in 64-99% of samples, depending on the source. In the subset where such terms are present, document-based datasets (FinePDFs and EssentialWeb) contain representational skews, such as references to \"male\" outnumbering those to \"female\", and mentions of \"White\" as the most frequent among ethnic identifiers (comprising 43-44% of ethnicity mentions). To mitigate these imbalances, we recommend considering evaluation techniques such as bias audits, fine-tuning with demographically balanced datasets, and mitigation strategies like counterfactual data augmentation to align with the desired model behavior. This evaluation used a 3,000-sample subset per dataset, identified as the optimal threshold for maximizing embedder accuracy.

During post-training, we generate synthetic data by distilling trajectories, solutions, and translations from strong teacher models and agent systems, often grounded in real tasks or documents and aggressively filtered for quality. For math, code, and science, we start from curated problem sets and use open source permissive models such as GPT-OSS-120B to produce step-by-step reasoning traces, candidate solutions, best-of-n selection traces, and verified CUDA kernels. For long-context and science, we build synthetic QA and reasoning data by retrieving passages from long documents, generating MCQ/OpenQA questions and answers, and paraphrasing them into multiple prompt/response formats to ensure diversity. Across all pipelines we stack automated verification—compilers, numerical checks, language identification—to ensure our data is high quality.

For all domains, we apply a unified data filtering pipeline to ensure that only high-quality, license-compliant, and verifiable samples are used for post-training. We first discard malformed examples using structural checks (e.g., missing tool definitions when tool calls are present). We then aggressively filter reasoning traces exhibiting pathological repetition, such as repeated n-grams within a sliding window or across the entire trajectory, which we found to be a strong indicator of malformed or low-quality reasoning. Finally, based on internal audits of synthetically generated datasets, we observed that some teacher models occasionally produce reasoning traces and final responses that implicitly align with specific political entities or promote nationalistic narratives. To mitigate this, we apply targeted keyword- and regex-based filters and remove all trajectories matching such behavior.

Alongside the model, we release our final pre-training and post-training data, as outlined in this section. For ease of analysis, there is a sample set that is ungated. For all remaining code, math and multilingual data, gating and approval is required, and the dataset is permissively licensed for model training purposes.

More details on the datasets and synthetic data generation methods can be found in the technical report NVIDIA Nemotron 3 Ultra.

For Detailed Dataset Information: Click here!

Base Pre-Training Corpus (Nemotron 3 Foundation)

The foundation of the model is trained on the Nemotron-3-Ultra corpus, comprising the following datasets from the Nemotron Pretraining Datasets collection:

\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n

tr:nth-child(odd)]:bg-surface-sunken\">

Dataset Collection	Token Counts	Description
Nemotron-CC-v2 & v2.1	9.1T	A massive collection of English web data filtered from Common Crawl, including 2.5T+ tokens of new organic, translated, and synthetically rephrased content.
Nemotron-CC-Code-v1	427.9B	High-quality code tokens extracted from Common Crawl using the Lynx + LLM pipeline to preserve structure and equations.
Nemotron-Pretraining-Code-v1 & v2 & v3	1.7T	Curated GitHub code references with multi-stage filtering, deduplication, and large-scale synthetic code data.
Nemotron-CC-Math-v1	133.3B	High-quality math pre-training dataset preserving LaTeX formatting and mathematical structures.
Nemotron-Pretraining-Specialized-v1 & v1.1 & v1.2 & Nemotron-Pretraining-SFT-v1	660.0B	Synthetic datasets targeting specialized domains such as STEM reasoning and scientific coding.
Nemotron-Pretraining-Legal-v1	4.3B	Synthetic datasets targeting the legal domain.

Public Datasets

tr:nth-child(odd)]:bg-surface-sunken\">

Dataset	Collection Period
GSM8K	4/23/2025
CC-NEWS	4/23/2025
Common Crawl	4/23/2025
Wikimedia	4/23/2025
Bespoke-Stratos-17k	4/23/2025
tigerbot-kaggle-leetcodesolutions-en-2k	4/23/2025
glaive-function-calling-v2	4/23/2025
APIGen Function-Calling	4/23/2025
LMSYS-Chat-1M	4/23/2025
Open Textbook Library - CC BY-SA & GNU subset and OpenStax - CC BY-SA subset	4/23/2025
Advanced Reasoning Benchmark, tigerbot-kaggle-leetcodesolutions-en-2k, PRM800K, and SciBench	4/23/2025
FineWeb-2	4/23/2025
Court Listener	Legacy Download
peS2o	Legacy Download
OpenWebMath	Legacy Download
BioRxiv	Legacy Download
PMC Open Access Subset	Legacy Download
OpenWebText2	Legacy Download
Stack Exchange Data Dump	Legacy Download
PubMed Abstracts	Legacy Download
NIH RePORTER ExPORTER Data Dictionary	Legacy Download
arXiv	Legacy Download
BigScience Workshop Datasets	Legacy Download
Pushshift Reddit data access	Legacy Download
SEC's Electronic Data Gathering, Analysis, and Retrieval (EDGAR)	Legacy Download
Advanced Mathematical Problem Solving	Legacy Download
MathPile	Legacy Download
NuminaMath CoT	Legacy Download
PMC Article	Legacy Download
FLAN	Legacy Download
Advanced Reasoning Benchmark	Legacy Download
SciBench	Legacy Download
WikiTableQuestions	Legacy Download
FinQA	Legacy Download
Riddles	Legacy Download
Problems in Elementary Mathematics for Home Study	Legacy Download
MedMCQA	Legacy Download
Cosmos QA	Legacy Download
MCTest	Legacy Download
AI2's Reasoning Challenge	Legacy Download
OpenBookQA	Legacy Download
MMLU Auxiliary Train	Legacy Download
social-chemestry-101	Legacy Download
Moral Stories	Legacy Download
The Common Pile v0.1	Legacy Download
FineMath	Legacy Download
MegaMath	Legacy Download
MegaMath	Legacy Download
MultiverseMathHard	10/2/2025
News Commentary	10/2/2025
Essential-Web	10/2/2025
finepdfs	10/2/2025
HotpotQA	10/2/2025
SQuAD2.0	10/2/2025
NLTK Words Lists	10/2/2025

Crawled and Scraped from Online Sources by NVIDIA

The English Common Crawl data was downloaded from the Common Crawl Foundation (see their FAQ for details on their crawling) and includes the snapshots CC-MAIN-2013-20 through CC-MAIN-2025-13. The data was subsequently deduplicated and filtered in various ways described in the Nemotron-CC paper. Additionally, we extracted data for fifteen languages from the following three Common Crawl snapshots: CC-MAIN-2024-51, CC-MAIN-2025-08, CC-MAIN-2025-18. The fifteen languages included were Arabic, Chinese, Danish, Dutch, French, German, Italian, Japanese, Korean, Polish, Portuguese, Russian, Spanish, Swedish, and Thai. As we did not have reliable multilingual model-based quality classifiers available, we applied just heuristic filtering instead—similar to what we did for lower quality English data in the Nemotron-CC pipeline, but selectively removing some filters for some languages that did not work well. Deduplication was done in the same way as for Nemotron-CC.

The GitHub Crawl was collected using the GitHub REST API and the Amazon S3 API. Each crawl was operated in accordance with the rate limits set by its respective source, either GitHub or S3. We collect raw source code and subsequently remove any having a license which does not exist in our permissive-license set (for additional details, refer to the technical report).

\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n

tr:nth-child(odd)]:bg-surface-sunken\">

Dataset	Modality	Dataset Size	Collection Period	Collecting Organisation
English Common Crawl	Text	3.36T	4/8/2025	NVIDIA Advanced Deep Learning Research
English Common Crawl 1.1	Text	Not disclosed	10/2/2025	NVIDIA Advanced Deep Learning Research
Multilingual Common Crawl	Text	812.7B	5/1/2025	NVIDIA Advanced Deep Learning Research
GitHub Crawl	Text	747.4B	4/29/2025	NVIDIA Advanced Deep Learning Research
GitHub Crawl 1.1	Text	172.7B	9/30/2025	NVIDIA Advanced Deep Learning Research

Private Non-publicly Accessible Datasets of Third Parties

\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n

tr:nth-child(odd)]:bg-surface-sunken\">

Dataset	Model(s) used
Global Regulation	Unknown
TAUS Translation Memory	Unknown
Scale HLE	Unknown
HackerRank Coding	Unknown
RL data for Search	Gemini 3; GPT-5

Private Non-publicly Accessible Datasets by NVIDIA

\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n

tr:nth-child(odd)]:bg-surface-sunken\">

Dataset	Model(s) used
Simple Minesweeper	Undisclosed
Simple Sudoku	Undisclosed
Multitool Typewriter Hard	Undisclosed
Machine Translation of News Commentary and TAUS Translation Memory	Undisclosed
Machine Translation of STEM -	Qwen2.5-14B-Instruct
Competitive Coding RL data from Nemotron Cascade	Undisclosed
Long context RL	Undisclosed
Single-step SWE RL for patch generation	Undisclosed
OpenHands SWE	Undisclosed

NVIDIA-Sourced Synthetic Datasets (Pre-Training)

tr:nth-child(odd)]:bg-surface-sunken\">

Dataset	Modality	Dataset Size	Seed Dataset	Model(s) used for generation
Nemotron-Pretraining-Fact-Seeking	Text	35.0B	FineWiki	Qwen3-30B-A3B-Instruct-2507
Nemotron-Pretraining-Legal	Text	4.3B	CommonPile (caselaw_access_project_filtered); California Code of Regulations; Judicial Ethics Opinions; GLOBALCIT; CUAD; Nemotron Personas; ToSDR Terms of Service Corpus; CodeHima/TOS_Dataset; ContractNLI; CaseHOLD; Code of Federal Regulations; Canadian Case Law (subsets that allow commercial use)	Qwen3-235B-A22B-Thinking-2507
Nemotron-Pretraining-Formal-Logic	Text	128M	Nemotron Personas	Qwen3-235B-A22B-Thinking-2507
Nemotron-Pretraining-Economics	Text	73.4M	-	Qwen3-235B-A22B-Thinking-2507
Nemotron-Pretraining-Multiple-Choice	Text	1.6B	MMLU Auxiliary Train	DeepSeek-V3; Qwen3-235B-A22B
Nemotron-Pretraining-Code-Concepts	Text	7.3B	-	gpt-oss-20b; gpt-oss-120b
Nemotron-Pretraining-Unconditional-Algorithmic	Text	196.5M	-	gpt-oss-120b; Qwen3-235B-A22B
More Synthetic Tasks from DeepSeek-V3 and Qwen3-235B-A22B	Text	1.1B	train splits of acp_bench; ai2_arc; babi; gsm8k; hendrycks_math; IFEval; MedText; mediqa_qa; mlqa; MMLU-Pro; mmlu-pro-plus; MMLU-ProX; nq_open; tinyGSM8k; truthful_qa; truthfulqa-multi; MATH-lighteval; mmlu; awesome-chatgpt-prompts; super_glue	DeepSeek v3; Qwen3-235B-A22B
Synthetic Tasks from DeepSeek-V3 and Qwen3-235B-A22B	Text	6.7B	train splits of Into the Unknown; AI2 ARC (AI2 Reasoning Challenge); BLiMP (Benchmark of Linguistic Minimal Pairs); CommonSenseQA; GLUE; HeadQA; Hendrycks Ethics; Memo Trap; modus-tollens; NeQA; pattern-matching-suppression; mastermind_24_mcq_random; mastermind_24_mcq_close; quote-repetition; redefine-math; Repetitive Algebra; sig-figs; MMLU-Pro; MC-TACO; MedConceptsQA; MMLU_dataset; OpenbooksQA; PIQA (Physical Interaction Question Answering); SocialIQA; SuperGLUE; tinyAI2_arc; tinyMMLU; tinyWinogrande; TruthfulQA; WebQuestions; Winogrande; GPQA; MBPP	DeepSeek v3; Qwen3-235B-A22B
Synthetic Art of Problem Solving from DeepSeek-R1	Text	40B	Art of Problem Solving; American Mathematics Competitions 8; American Mathematics Competitions 10;	DeepSeek-R1
Synthetic Moral Stories and Social Chemistry from Qwen3-235B-A22B-Thinking-2507 and Mixtral-8x22B-v0.1	Text	15.2M	social-chemestry-101; Moral Stories	Qwen3-235B-A22B-Thinking-2507; Mixtral-8x22B-v0.1
Synthetic Moral Stories and Social Chemistry from Mixtral-8x22B-v0.1	Text	327M	social-chemestry-101; Moral Stories	Mixtral-8x22B-v0.1
Synthetic Social Sciences seeded with OpenStax from DeepSeek-V3, Mixtral-8x22B-v0.1, and Qwen2.5-72B	Text	83.6M	OpenStax - CC BY-SA subset	DeepSeek-V3; Mixtral-8x22B-v0.1; Qwen2.5-72B
Synthetic Health Sciences seeded with OpenStax from DeepSeek-V3, Mixtral-8x22B-v0.1, and Qwen2.5-72B	Text	9.7M	OpenStax - CC BY-SA subset	DeepSeek-V3; Mixtral-8x22B-v0.1; Qwen2.5-72B
Synthetic STEM seeded with OpenStax, Open Textbook Library, and GSM8K from DeepSeek-R1, DeepSeek-V3, DeepSeek-V3-0324, and Qwen2.5-72B	Text	175M	OpenStax - CC BY-SA subset; GSM8K; Open Textbook Library - CC BY-SA & GNU subset	DeepSeek-R1, DeepSeek-V3; DeepSeek-V3-0324; Qwen2.5-72B
Nemotron-PrismMath	Text	4.6B	Big-Math-RL-Verified; OpenR1-Math-220k	Qwen2.5-0.5B-instruct, Qwen2.5-72B-Instruct; DeepSeek-R1-Distill-Qwen-32B
Synthetic Question Answering Data from Papers and Permissible Books from Qwen2.5-72B-Instruct	Text	350M	arXiv; National Institutes of Health ExPorter; BioRxiv; PMC Article; USPTO Backgrounds; peS2o; Global Regulation; CORE; PG-19; DOAB CC BY & CC BY-SA subset; NDLTD ETD resources	Qwen2.5-72B-Instruct
Synthetic Rephrased Math Data from Common Crawl from phi-4	Text	73B	Common Crawl	phi-4
Synthetic Math Data from Common Crawl 4plus	Text	52.3B	Common Crawl	phi-4
Synthetic Math Data from Common Crawl 3	Text	80.9B	Common Crawl	phi-4
Synthetic AGIEval seeded with AQUA-RAT, LogiQA, and AR-LSAT from DeepSeek-V3 and DeepSeek-V3-0324	Text	4.0B	AQUA-RAT; LogiQA; AR-LSAT	DeepSeek-V3; DeepSeek-V3-0324
Synthetic AGIEval seeded with AQUA-RAT, LogiQA, and AR-LSAT from Qwen3-30B-A3B	Text	4.2B	AQUA-RAT; LogiQA; AR-LSAT	Qwen3-30B-A3B
Synthetic Art of Problem Solving from Qwen2.5-32B-Instruct, Qwen2.5-Math-72B, Qwen2.5-Math-7B, and Qwen2.5-72B-Instruct	Text	Undisclosed	Art of Problem Solving; American Mathematics Competitions 8; American Mathematics Competitions 10; GSM8K; PRM800K	Qwen2.5-32B-Instruct; Qwen2.5-Math-72B; Qwen2.5-Math-7B; Qwen2.5-72B-Instruct
Synthetic MMLU Auxiliary Train from DeepSeek-R1	Text	0.5B	MMLU Auxiliary Train	DeepSeek-R1
Synthetic Long Context Continued Post-Training Data from Papers and Permissible Books from Qwen2.5-72B-Instruct	Text	Undisclosed	arXiv; National Institutes of Health ExPorter; BioRxiv; PMC Article; USPTO Backgrounds; peS2o; Global Regulation; CORE; PG-19; DOAB CC BY & CC BY-SA subset; NDLTD ETD resources	Qwen2.5-72B-Instruct
Synthetic Common Crawl from Qwen3-30B-A3B and Mistral-Nemo-12B-Instruct	Text	415.8B	Common Crawl	Qwen3-30B-A3B; Mistral-NeMo-12B-Instruct
Synthetic Multilingual Data from Common Crawl from Qwen3-30B-A3B	Text	Undisclosed	Common Crawl	Qwen3-30B-A3B
Synthetic Multilingual Data from Wikimedia from Qwen3-30B-A3B	Text	Undisclosed	Wikimedia	Qwen3-30B-A3B
Synthetic Math Data from Wikimedia from Nemotron-4-340B-Instruct	Text	Undisclosed	-	Nemotron-4-340B-Instruct
Synthetic Common Crawl Code from phi-4	Text	427.9B	Common Crawl	phi-4
Synthetic Scientific Coding from Qwen3-235B-A22B	Text	1.2B	Wikimedia	Qwen3-235B-A22B
Tool Calling Data	Text	26.2B		Qwen3-235B-A22B-2507; gpt-oss-120b
Synthetic Essential-Web from QwQ-32B	Text	28.1B	Essential-Web	QwQ-32B
Translated Synthetic Crawl	Text	389.9B	Common Crawl	Qwen3-30B-A3B
Translated Synthetic Wikipedia	Text	7.9B	Wikimedia	Qwen3-30B-A3B
Synthetic Long Context from Qwen3-235B-A22B-Instruct-2507	Text	Undisclosed	CORE; PG-19; DOAB CC BY & CC BY-SA subset; NDLTD ETD resources	Qwen3-235B-A22B-Instruct-2507
Synthetic Search STEM OPENQ from DeepSeek-R1-0528	Text	Undisclosed	-	DeepSeek-R1-0528
Synthetic MCQ from Qwen2.5-32B-Instruct and DeepSeek-R1-0528	Text	Undisclosed	-	Qwen2.5-32B-Instruct; DeepSeek-R1-0528
Synthetic Offline Search MCQA HLE from DeepSeek-R1-0528	Text	Undisclosed	-	DeepSeek-R1-0528
Synthetic Offline Search MCQA GPQA from Qwen3-235B-A22B and DeepSeek-R1-0528	Text	Undisclosed	-	Qwen3-235B-A22B; DeepSeek-R1-0528
Synthetic Human Preference from QwQ-32B, Qwen3-30B-A3B, Qwen3-235B-A22B, Qwen3-235B-A22B-Instruct-2507, Mistral-Small-3.1-24B-Instruct-2503, Mistral-Small-3.2-24B-Instruct-2506, MiniMax-M1-80k, MiniMax-M1-40k, Kimi-K2-Instruct, DeepSeek-V3-0324, DeepSeek-R1-0528	Text	Undisclosed	-	QwQ-32B; Qwen3-30B-A3B; Qwen3-235B-A22B; Qwen3-235B-A22B-Instruct-2507; Mistral-Small-3.1-24B-Instruct-2503; Mistral-Small-3.2-24B-Instruct-2506; MiniMax-M1-80k; MiniMax-M1-40k; Kimi-K2-Instruct; DeepSeek-V3-0324; DeepSeek-R1-0528
Synthetic Code from Qwen3-32B	Text	Undisclosed	English Common Crawl; English Common Crawl 1.1	Qwen3-32B
Synthetic OpenCodeReasoning from DeepSeek-R1	Text	Undisclosed	OpenCodeReasoning	DeepSeek-R1
Synthetic LIMO from DeepSeek-R1-0528	Text	Undisclosed	LIMO	DeepSeek-R1-0528
Synthetic SCP from DeepSeek-R1-0528	Text	Undisclosed	SCP-116K	DeepSeek-R1-0528
Synthetic Stack Exchange from DeepSeek-R1-0528	Text	Undisclosed	Stack Exchange	DeepSeek-R1-0528
Synthetic Common Crawl from Qwen3-30B-A3B	Text	Undisclosed	Common Crawl	Qwen3-30B-A3B
Synthetic Wikipedia from Qwen3-30B-A3B	Text	Undisclosed	Wikimedia	Qwen3-30B-A3B
Synthetic Essential-Web from Qwen3-30B-A3B and Qwen3-235B-A22B-Thinking-2507	Text	Undisclosed	Essential-Web	Qwen3-30B-A3B; Qwen3-235B-A22B-Thinking-2507
Synthetic Textbook Math from Qwen3-30B-A3B, Qwen3-235B-A22B, phi-4	Text	Undisclosed	Common Crawl; FineMath	Qwen3-30B-A3B; Qwen3-235B-A22B; phi-4
Synthetic Math and Code from DeepSeek-R1 and DeepSeek-R1-0528	Text	Undisclosed	Magicoder-Evol-Instruct-110K; opc-sft-stage2; TACO; OpenCodeReasoning; OpenMathReasoning; NuminaMath CoT	DeepSeek-R1; DeepSeek-R1-0528

\n
\n

NVIDIA-Sourced Synthetic Datasets (Post-Training)

\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n

tr:nth-child(odd)]:bg-surface-sunken\">

Dataset	Modality	Dataset Size	Seed Dataset	Model(s) used for generation
Synthetic Competitive MATH Proofs from DeepSeek-V4-Pro	Text	Undisclosed	[AMC 8 Problems and Solutions, AMC 10 Problems and Solution, and AIME Problems and Solutions]	[deepseek-ai/DeepSeek-V4-Pro]
Synthetic Hermes Agent Reasoning Traces	Text	Undisclosed	[lambda/hermes-agent-reasoning-traces]	[hermes-agent-generator]
Synthetic Competitive Coding from DeepSeek-V4-Pro	Text	Undisclosed	[NVCompetitiveCodingV1]	[deepseek-ai/DeepSeek-V4-Pro]
Synthetic Competitive Science Reasoning from DeepSeek-V4-Pro	Text	Undisclosed	[AMC 8 Problems and Solutions, AMC 10 Problems and Solution, and AIME Problems and Solutions]; [EssentialAI/essential-web-v1.0]; [cdquestions.com]; [Pile-FreeLaw]; [Vedantu]; [askfilo]; [doubtnut]; [ICHO-IPH0 Dataset]; [LIMO dataset (Less is More for Reasoning)]; [AAPT]; [ChemData 700K]; [oMeBench]; [Flavor Analysis and Recognition Transformer]; [ChemCoTBench]; [Llama Nemotron Dataset]	[deepseek-ai/DeepSeek-V4-Pro]
Synthetic Competitive MATH CoT and TIR from Nemotron 5.5	Text	Undisclosed	[Pile-FreeLaw]; [AMC 8 Problems and Solutions, AMC 10 Problems and Solution, and AIME Problems and Solutions]	[Nemotron 5.5]
Vendor Terminal Bench-like Tasks from Mercor	Text	Undisclosed	[Terminal bench like tasks curated by the vendor]	[Undisclosed - purchased dataset]
Turing Math Data Pack	Text	Undisclosed	[Turing Math Data Pack dataset]	[Undisclosed - purchased dataset]
Synthetic Holdout, Skywork, DAPO, and Turing Math from GPT-5.5	Text	Undisclosed	[DocQA-RL-1.6K]; [DAPO-Math-17k]	[GPT-5.5]
Synthetic Long Context RL from QwenLong L1 and DocQA-RL-1.6K	Text	Undisclosed	[DocQA-RL-1.6K]	Undisclosed
Synthetic Competitive Coding Gym Tasks	Text	Undisclosed	[NVCompetitiveCodingV1.1]	Undisclosed
Synthetic Finance SEC Search Agent from GPT-OSS-120B and Qwen3	Text	Undisclosed	[SEC filings from sec.gov]	[GPT-OSS-120B]; [Qwen3-235B-A22B-Instruct]; [Qwen3-4B-Instruct]
Synthetic Structured Outputs from Qwen3-30B-A3B-Instruct-2507, Qwen3-30B-A3B-Thinking-2507, Qwen3-235B-A22B-Instruct-2507, and Qwen3-235B-A22B-Thinking-2507	Text	Undisclosed	[Nemotron-RL-agent-structured-outputs-v1]	[Qwen3-30B-A3B-Instruct-2507]; [Qwen3-235B-A22B-Instruct-2507]
Synthetic Long Context Equivalence Rule from Qwen3-235B-A22B-Thinking-2507 and DeepSeek-R1	Text	Undisclosed	[Long-context SFT data]	[Qwen/Qwen3-235B-A22B-Thinking-2507]; [Deepseek-ai/DeepSeek-R1]
Synthetic Science RL Data Blend from Qwen2.5-32B	Text	Undisclosed	[doubtnut]; [Pile-FreeLaw]; [Llama Nemotron Dataset]; [askfilo]; [EssentialAI/essential-web-v1.0]; [Vedantu]; [auxiliary_train]; [cdquestions.com]; [AMC 8 Problems and Solutions, AMC 10 Problems and Solution, and AIME Problems and Solutions]; [AAPT]; [ICHO-IPH0 Dataset]; [LIMO dataset (Less is More for Reasoning)]	[Qwen2.5-32B]
Synthetic Abstention Data from Nemotron Super v3	Text	Undisclosed	[Go abstention Dataset]	[nvidia/nvidia/nemotron-3-super-v3]
Synthetic Chemistry Data from Nemotron Super v3	Text	Undisclosed	[ChemData 700K]	[nvidia/nvidia/nemotron-3-super-v3]
Synthetic Structured Outputs from Qwen3-30B-A3B-Instruct-2507, Qwen3-30B-A3B-Thinking-2507, Qwen3-235B-A22B-Instruct-2507, and Qwen3-235B-A22B-Thinking-2507	Text	Undisclosed	[In-house data]	[GPT OSS 120B - Apache 2.0]
Synthetic Tool Call Schema for RL	Text	Undisclosed	[In-house data]	[GPT OSS 120B - Apache 2.0]
Synthetic Freeform Text Formatting from GPT-OSS-120B	Text	Undisclosed	[In-house data]	[GPT OSS 120B - Apache 2.0]
Synthetic Citation Formatting from GPT-OSS-120B	Text	Undisclosed	[In-house data]	[GPT OSS 120B - Apache 2.0]
Droid Harness Pivot Vendor Data	Text	Undisclosed	[Droid Harness Pivot vendor data]	Undisclosed
Synthetic HotpotQA Training Data from Qwen3-235B	Text	Undisclosed	[HotpotQA]	[Qwen3-235B]
Synthetic Natural Language Math Proofs from Nemotron 5.5	Text	Undisclosed	[AMC8, AMC10, and AIME problem sets hosted on Art of Problem Solving]; [Pile-StackExchange]	[Nemotron 5.5]
Synthetic Stack Overflow OpenQ	Text	Undisclosed	[Pile-FreeLaw]	Undisclosed
Chemistry Ether0 Vendor Data	Text	Undisclosed	[Chemistry ether0 vendor data]	Undisclosed
Synthetic Litmus-Bench Chemistry from ChEMBL	Text	Undisclosed	[ChEMBL]; [Nemo Gym RL dataset generated from ChEMBL with RDKit]	Undisclosed
Synthetic ZINC Chemistry from Nemotron Super v3	Text	Undisclosed	[ZINC]	[Nemotron Super v3]
ARC-AGI Gym Environment	Text	Undisclosed	[ARC-AGI-2]	[ARC-AGI-2]
Synthetic Agentic Search Tool-Use from DeepSeek-V3.2	Text	Undisclosed	[Mercor Data]	[DeepSeek-V3.2]
Synthetic Text-To-SQL	Text	Undisclosed	[In-house Text-to-SQL data]	[gpt-oss-120b]
Dialog Memory Vendor Data	Text	Undisclosed	[Patronus external vendor agreement]	Undisclosed
Synthetic Indirect Prompt Injection from Nemotron Super v3 and Qwen3-Next-80B-A3B-Instruct	Text	Undisclosed	[In-house indirect prompt injection data]	[nvidia/nemotron-3-super-v3, qwen/qwen3-next-80b-a3b-instruct.]
Synthetic Malicious Code and Agentic Security	Text	Undisclosed	[In-house malicious-code / agentic-security data]	Undisclosed
Synthetic Single-Step SWE Patch Selection	Text	Undisclosed	[SWE-Gym Dataset]; [SWE Bench Verified Benchmark]	[ground truth and task checks]
Synthetic Natural Language Math Final Answers from Nemotron 5.5	Text	Undisclosed	[AMC8, AMC10, and AIME problem sets hosted on Art of Problem Solving]; [Pile-StackExchange]	[nemotron 5.5]
Synthetic Simple Math Prompts for Token Efficiency	Text	Undisclosed	[In-house simple math prompts]	Undisclosed
Synthetic Abstention Data from Nemotron Super v3	Text	Undisclosed	[CRAG]	[nvidia/nvidia/nemotron-3-super-v3]
Synthetic Agentless SWE	Text	242,536	[SWE-Rebench-V2]; [SWEbench Training Set]; [R2E-Gym/R2E-Gym-Subset]; [SWE-Gym/SWE-Gym]; [SWE-Rebench]	[openai/gpt-oss-120b]
Synthetic Agentic CUDA Traces from GLM-4.7	Text	2,276	[Internal CUDA task data]	[GLM-4.7]
Synthetic Math Proofs from DeepSeek-V3.2-Speciale	Text	820,772	[Nemotron-Math-Proofs-v1]	[SDG: DeepSeek-V3.2-Speciale]; [Filter: proof validation]
Synthetic Multilingual SFT from DeepSeek-V3	Text	1,245,284	[Nano v3 SFT data]	[DeepSeek-V3]
Synthetic Agentic Code from gpt-oss-120b	Text	109,086	[NVAgenticCLIPrompts-v1]; [NVAgenticSkills-v1]; [NVAgenticCLIMultiTurnPrompts-v1]	[openai/gpt-oss-120b]
Synthetic Agentic CLI and Web Skills from gpt-oss-120b	Text	27,418	[NVAgenticCLIPrompts-v1]; [NVAgenticSkills-v1]; [NVAgenticCLIMultiTurnPrompts-v1]; [NVAgenticCLIPrompts-Web-v1]	[openai/gpt-oss-120b]
Synthetic Agentic Coding from gpt-oss-120b	Text	160,531	[NVAgenticCLIPrompts-v1]; [NVAgenticSkills-v1]; [NVAgenticCLIMultiTurnPrompts-v1]; [NVAgenticCLIPrompts-Web-v1]	[openai/gpt-oss-120b]
Synthetic OpenCode Agentic Tasks from gpt-oss-120b	Text	614,773	[NVAgenticCLIPrompts-v1]; [NVAgenticSkills-v1]; [NVAgenticCLIMultiTurnPrompts-v1]; [NVAgenticCLIPrompts-Web-v1]	[openai/gpt-oss-120b]
Synthetic ARC-AGI Ultra Data	Text	192,016	[ARC-AGI-2]; [arc dataset collection]	[ARC-AGI-2]
Synthetic LiveCodeBench TIR from DeepSeek-R1-0528	Text	1,283,398	[Nemotron-X training datasets]	[DeepSeek-R1-0528]
Synthetic Verilog and SystemVerilog Code from DeepSeek-R1-0528 and GPT-OSS-120B	Text	1,233,247	[Verilog/SystemVerilog seed code]	[SDR: DeepSeek R1 0528 and GPT-OSS-120B]; [Filtering: Claude 4 Sonnet]
Synthetic Aider Python Tasks from DeepSeek-R1-0528	Text	236,099	[Exercism (GitHub Python)]	[Deepseek R1 0528]
Synthetic Chat Reasoning-Off Data from GLM-5	Text	646,738	[lmarena-ai/repochat-arena-preference-4k user prompts]	[Multi-turn conversations generated by GLM-5 with best-of-4 selection via Qwen3-Nemotron-235B-A22B-GenRM:]
Synthetic Chat Reasoning-On Data from GLM-5	Text	644,286	[lmarena-ai/repochat-arena-preference-4k user prompts]; [lmarena-ai/arena-expert-5k user prompts]; [lmarena-ai/arena-human-preference-55k user prompts]; [lmarena-ai/arena-human-preference-100k user prompts]; [lmarena-ai/arena-human-preference-140k user prompts]	[Multi-turn conversations generated by GLM-5 with best-of-4 selection via Qwen3-Nemotron-235B-A22B-GenRM:]
Synthetic Multilingual Safety from Riva-Translate-4B-Instruct-v1.1	Text	132,067	[Safety SFT Data: Ultra]	[nvidia/Riva-Translate-4B-Instruct-v1.1]
Synthetic Science Reasoning Effort Medium	Text	502,722	[science-reasoning-effort-medium-v0]	Undisclosed
Synthetic Telecom Tool-Use Trajectories from gpt-oss-120b	Text	12,455	[Existing Tau2 telecom trajectories originally generated with DeepSeek V3.2]	[gpt-oss-120b]
Synthetic Terminal Bench Data from OpenReasoningv2	Text	Undisclosed	[OpenCodeReasoningv2]; [OpenMathReasoning]; [nemo-swe-bench-repos]; [SWE-Rebench]; [SWE-Fixer-110K]	[OpenReasoningv2]
Synthetic Tulu Instruction Following from DeepSeek-R1-0528	Text	105,361	[Nemotron-X training datasets]	[DeepSeek-R1-0528]
Synthetic SWE Unverified	Text	Undisclosed	[NVAgenticCLIPrompts-v1]; [NVAgenticSkills-v1]; [NVAgenticCLIMultiTurnPrompts-v1]; [NVAgenticCLIPrompts-Web-v1]	[gpt-oss-120b]
Synthetic Instruction Following from gpt-oss-120b	Text	151,988	[IFEval]; [IFEvalG]	[gpt-oss-120b]
Synthetic Identity Data from Qwen3-Next-80B-A3B-Instruct and Qwen3-235B-A22B-Instruct-2507	Text	25,992	[Hand-written prompts]	[Qwen3-Next-80B-A3B-Instruct]; [Qwen3-235B-A22B-Instruct-2507]
Synthetic Terminus Ultra Agentic Reasoning Blend	Text	96,881	[ARC-AGI-2]; [OpenCodeReasoningv2]; [OpenMathReasoning]; [SWE-Fixer-110K]; [SWE-Rebench]; [SWE-Smith]	[DeepSeek-V3.2]; [Qwen3-235B-A22B-Thinking-2507]; [Ring-1T]; [Kimi-K2.5]; [GLM-4.7-FP8]; [Qwen3-Next-80B-A3B-Thinking]; [gpt-oss-120b]; [Ministral-3-14B-Reasoning-2512]; [LM-4.5-Air-FP8]
Synthetic STEM from Qwen3-235B-A22B-Thinking-2507	Text	1,174,694	[IChO-IPhO-RL-v2]; [Physics-Big Dataset]	Undisclosed
Translation Data from TAUS	Text	1,618,055	[TAUS proprietary dataset]	Undisclosed
Synthetic Art of Problem Solving and Stack Exchange from gpt-oss-120b, Qwen2.5-32B-Instruct, and Goedel-Prover-V2-32B	Text	860,469	[Nemotron-Math-Proofs-v1]	[Goedel-Prover-V2-32B]
Synthetic Art of Problem Solving and Stack Exchange from gpt-oss-120b, Qwen2.5-32B-Instruct, and Goedel-Prover-V2-32B	Text	1,201,815	[Upstream released math dataset]; [AoPS]; [StackOverflow / StackExchange]	[gpt-oss-120b]
Synthetic Art of Problem Solving and Stack Exchange from gpt-oss-120b, Qwen2.5-32B-Instruct, and Goedel-Prover-V2-32B	Text	1,296,676	[Upstream released math dataset]; [AoPS]; [StackOverflow / StackExchange]	[gpt-oss-120b]
Synthetic Instruction Following for RL	Text	Undisclosed	[WildChat-1M]; [LMSYS-340B-Eval Dataset]; [LMSYS-Chat-1M Prompts]; [IFEval]; [IFEvalG]	[Qwen/Qwen3-235B-A22B-Thinking-2507]; [gpt-oss-120b]; [Qwen3-235B-A22B-Instruct-2507]
Synthetic Instruction Following for RL	Text	Undisclosed	[WildChat-1M]; [LMSYS-340B-Eval Dataset]; [LMSYS-Chat-1M Prompts]; [IFEval]; [IFEvalG]	[Qwen/Qwen3-235B-A22B-Thinking-2507]; [gpt-oss-120b]; [Qwen3-235B-A22B-Instruct-2507]
Synthetic Multilingual Science and Code data from DeepSeek-R1, DeepSeek-R1-0528, Qwen2.5-32B-Instruct, and Qwen3-235B-A22B, translated with Qwen2.5-32B-Instruct and Qwen2.5-14B-Instruct	Text	Undisclosed	[Nano-V3 SFT Data (without tool call)]	[Qwen/Qwen2.5-14B-Instruct]; [Qwen/Qwen3-4B-Thinking-2507]
Synthetic Search Graph Walk	Text	6,977	[Wikidata / Wikipedia KnowledgeBase]	[MiniMaxAI/MiniMax-M2]
Synthetic Agentic Diverse Domains	Text	281,537	[Handwritten prompts (synthetic; no external seed data used)]	[SDG model: deepseek-ai/DeepSeek-V3.2, deepseek-ai/DeepSeek-R1-0528, Qwen/Qwen3-235B-A22B-Thinking-2507, Qwen/Qwen3-32B]; [Filtering model: openai/gpt-oss-120b, Qwen/Qwen3-32B, Qwen/Qwen3-235B-A22B-Instruct-2507]
Synthetic Long Context from Qwen3-235B-A22B-Instruct-2507	Text	65,608	[Long-context SFT seed blend (pre-training blend + nano-v1 post-training data)]	[Qwen/Qwen3-235B-A22B-Thinking-2507 and deepseek-ai/DeepSeek-R1]
Synthetic Agentless SWE	Text	209,976	[SWE-Bench-Train]; [SWE-Fixer-Train]; [SWE-reBench]; [SWE-Smith]	[deepseek-ai/DeepSeek-R1-0528]
Synthetic Nemotron Math SFT from DeepSeek-V3.2-Speciale	Text	1,900,553	[Nemotron-Math-v2 (AOPS and StackExchange-math problems)]	[DeepSeek-V3.2-Speciale]
Synthetic Nemotron Math TIR from DeepSeek-V3.2	Text	1,789,258	[Nemotron-Math-v2 (AOPS and StackExchange-math problems)]	[DeepSeek-V3.2]
Synthetic SWE Unverified	Text	27,911	[NVAgenticCLIPrompts-v1]; [NVAgenticSkills-v1]	[gpt-oss-120b]
Synthetic SWE Unverified	Text	28,116	[NVAgenticCLIPrompts-v1]; [NVAgenticSkills-v1]	[Qwen3-Coder-480B-A35B-Instruct]
Synthetic NemoCascade OCR Distillation from gpt-oss-120b	Text	682,864	[Nemotron-X training datasets]	[gpt-oss-120b]
Synthetic SWE Unverified	Text	26,865	[NVAgenticCLIPrompts-v1]; [NVAgenticSkills-v1]	[gpt-oss-120b]; [Qwen/Qwen3-Coder-480B-A35B-Instruct]; [GLM-4.7-Flash]
Synthetic CUDA 100k	Text	93,086	[KernelBook]; [HuggingFace Transformers]; [FlashInfer]	[gpt-oss-120b]; [DeepSeek-R1-0528]
Synthetic Science MCQ and QA Diversity from GPT-OSS and Kimi-K2	Text	30,358	[doubtnut]; [Pile-FreeLaw]; [Llama Nemotron Dataset]; [askfilo]; [EssentialAI/essential-web-v1.0]; [Vedantu]; [auxiliary_train]; [cdquestions.com]; [AMC 8 Problems and Solutions, AMC 10 Problems and Solution, and AIME Problems and Solutions]; [AAPT]; [ICHO-IPH0 Dataset]; [LIMO dataset (Less is More for Reasoning)]	[GPT-OSS]; [Kimi-K2]
Synthetic Science HLE with Python from GPT-OSS and Kimi-K2	Text	85,184	[doubtnut]; [Pile-FreeLaw]; [Llama Nemotron Dataset]; [askfilo]; [EssentialAI/essential-web-v1.0]; [Vedantu]; [auxiliary_train]; [cdquestions.com]; [AMC 8 Problems and Solutions, AMC 10 Problems and Solution, and AIME Problems and Solutions]; [AAPT]; [ICHO-IPH0 Dataset]; [LIMO dataset (Less is More for Reasoning)]	[GPT-OSS]; [Kimi-K2]
Synthetic Science Search and Python from GPT-OSS and Kimi-K2	Text	6,179	[doubtnut]; [Pile-FreeLaw]; [Llama Nemotron Dataset]; [askfilo]; [EssentialAI/essential-web-v1.0]; [Vedantu]; [auxiliary_train]; [cdquestions.com]; [AMC 8 Problems and Solutions, AMC 10 Problems and Solution, and AIME Problems and Solutions]; [AAPT]; [ICHO-IPH0 Dataset]; [LIMO dataset (Less is More for Reasoning)]	[GPT-OSS]; [Kimi-K2]
Synthetic Science Search from GPT-OSS and Kimi-K2	Text	32,554	[doubtnut]; [Pile-FreeLaw]; [Llama Nemotron Dataset]; [askfilo]; [EssentialAI/essential-web-v1.0]; [Vedantu]; [auxiliary_train]; [cdquestions.com]; [AMC 8 Problems and Solutions, AMC 10 Problems and Solution, and AIME Problems and Solutions]; [AAPT]; [ICHO-IPH0 Dataset]; [LIMO dataset (Less is More for Reasoning)]	[GPT-OSS]; [Kimi-K2]
Synthetic Finance Reasoning from GPT-OSS-120B and Qwen3-235B-A22B-Instruct-2507	Text	326,700	[_SEC filings]	[GPT-OSS-120B, Qwen3-235B-A22B-Instruct-2507]
Synthetic Science Diversity MCQ from GPT-OSS and Kimi-K2	Text	532,942	[doubtnut]; [Pile-FreeLaw]; [Llama Nemotron Dataset]; [askfilo]; [EssentialAI/essential-web-v1.0]; [Vedantu]; [auxiliary_train]; [cdquestions.com]; [AMC 8 Problems and Solutions, AMC 10 Problems and Solution, and AIME Problems and Solutions]; [AAPT]; [ICHO-IPH0 Dataset]; [LIMO dataset (Less is More for Reasoning)]	[GPT-OSS]; [Kimi-K2]
Synthetic Science Diversity OpenQ from GPT-OSS and Kimi-K2	Text	131,045	[doubtnut]; [Pile-FreeLaw]; [Llama Nemotron Dataset]; [askfilo]; [EssentialAI/essential-web-v1.0]; [Vedantu]; [auxiliary_train]; [cdquestions.com]; [AMC 8 Problems and Solutions, AMC 10 Problems and Solution, and AIME Problems and Solutions]; [AAPT]; [ICHO-IPH0 Dataset]; [LIMO dataset (Less is More for Reasoning)]	[GPT-OSS]; [Kimi-K2]
Synthetic Science Reasoning No-Tool from GPT-OSS and Kimi-K2	Text	2,085,600	[doubtnut]; [Pile-FreeLaw]; [Llama Nemotron Dataset]; [askfilo]; [EssentialAI/essential-web-v1.0]; [Vedantu]; [auxiliary_train]; [cdquestions.com]; [AMC 8 Problems and Solutions, AMC 10 Problems and Solution, and AIME Problems and Solutions]; [AAPT]; [ICHO-IPH0 Dataset]; [LIMO dataset (Less is More for Reasoning)]	[GPT-OSS]; [Kimi-K2]
Synthetic Long Context from Qwen3-235B-A22B-Instruct-2507	Text	62,333	[Long-context SFT data: lc_nothink 256k]	[Qwen/Qwen3-235B-A22B-Thinking-2507 and deepseek-ai/DeepSeek-R1]
Synthetic Long Context from Qwen3-235B-A22B-Instruct-2507	Text	49,698	[Long-context SFT data: MRCR 200k]	[Qwen/Qwen3-235B-A22B-Thinking-2507 and deepseek-ai/DeepSeek-R1]
Synthetic Text-To-SQL	Text	96,564	[Undisclosed - no seed data listed]	[gpt-oss-120b]
Synthetic Long Context from Qwen3-235B-A22B-Instruct-2507	Text	397,538	[Long-context SFT data: RULER 256k]	[Qwen/Qwen3-235B-A22B-Thinking-2507 and deepseek-ai/DeepSeek-R1]
Synthetic SWE Unverified	Text	27,960	[NVAgenticCLIPrompts-v1]; [NVAgenticSkills-v1]	[gpt-oss-120b]; [Qwen/Qwen3-Coder-480B-A35B-Instruct]; [GLM-4.7-Flash]
Synthetic SWE Unverified	Text	24,632	[NVAgenticCLIPrompts-v1]; [NVAgenticSkills-v1]	[gpt-oss-120b]; [Qwen/Qwen3-Coder-480B-A35B-Instruct]; [GLM-4.7-Flash]
Synthetic Long Context from Qwen3-235B-A22B-Instruct-2507	Text	49,902	[Long-context SFT data]	[Qwen/Qwen3-235B-A22B-Thinking-2507 and deepseek-ai/DeepSeek-R1]
Synthetic Tool Call Schema for RL	Text	469,983	[UltraTool]; [ToolEyes]; [AutoTools]; [API-Bank]; [Nemotron-Personas-USA]; [Salesforce xLAM function-calling]; [Glaive function-calling-v2]; [Agent-Ark/Toucan-1.5M]	[DeepSeek-V3.2]; [GLM-4.6]; [gpt-oss-120b]; [Kimi-K2-Instruct]
Synthetic Tool Call Schema for RL	Text	707,967	[UltraTool]; [ToolEyes]; [AutoTools]; [API-Bank]; [Nemotron-Personas-USA]; [Salesforce xLAM function-calling]; [Glaive function-calling-v2]; [Agent-Ark/Toucan-1.5M]	[DeepSeek-V3.2]; [GLM-4.6]; [gpt-oss-120b]; [Kimi-K2-Instruct]
Synthetic Long Context from Qwen3-235B-A22B-Instruct-2507	Text	52,630	[AALCR seed blend: SEC Filings]; [CC]; [Wikipedia]; [FinePDFs]; [ArXiv]; [Pile-NIH ExPorter]; [BioRxiv]; [PMC Article]; [USPTO Backgrounds]; [peS20]; [Global Regulations]; [CORE]; [Gutenberg (PG-19)]; [DOAB CC-BY]; [NDLTD]; [Amps]; [StackExchange]; [MathPile]; [Numinas]	[Qwen3-30B-A3B]
Synthetic Safety from gemma-3-4b-it, Nemotron-Nano-9B-v2, and gpt-oss-120b	Text	44,091	[Safety SFT Data]	[google/gemma-3-4b-it]; [Nemotron-Nano-9B-v2]; [gpt-oss-120b]

Language Distribution in Post-Training

For our post-training recipe, we focused on the following languages in addition to English: French, Spanish, Italian, German, Japanese, Korean, Hindi, Korean, Brazilian Portuguese, and Chinese.

Those languages were represented in the form of multilingual reasoning and translation tasks.

The following table depicts our sample distribution.

\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n

tr:nth-child(odd)]:bg-surface-sunken\">

Language	Size
English	8.6M
Italian	138k
German	138k
Spanish	138k
French	138k
Japanese	138k
Chinese	138k
Hindi	138k
Korean	138k
Brazilian Portuguese	138k

Evaluation Datasets:

Data Collection Method by dataset

Hybrid: Automated, Human, Synthetic

Labeling Method by dataset

Hybrid: Automated, Human, Synthetic

Properties: This corpus comprises a mix of high-quality standard benchmarks and test suites for modern agentic AI as outlined in the benchmark section of the model card.

Testing Datasets:

Data Collection Method by dataset

Hybrid: Automated, Human, Synthetic

Labeling Method by dataset

Hybrid: Automated, Human, Synthetic

Properties: This corpus comprises a mix of high-quality standard benchmarks and test suites for modern agentic AI as outlined in the benchmark section of the model card.

Inference

Acceleration Engine: PyTorch
Test Hardware:\n
- NVIDIA Hopper\n
  - H100
  - H200
  \n
- NVIDIA Grace Blackwell\n
  - GB200
  - GB300
  \n
- NVIDIA Blackwell\n
  - B200
  - B300
  \n
\n

Ethical Considerations

NVIDIA believes Trustworthy AI is a shared responsibility and we have established policies and practices to enable development for a wide array of AI applications. When downloaded or used in accordance with our terms of service, developers should work with their internal model team to ensure this model meets requirements for the relevant industry and use case and addresses unforeseen product misuse.

We advise against circumvention of any provided safety guardrails contained in the Model without a substantially similar guardrail appropriate for your use case. For more details: Safety and Explainability Subcards.

For more detailed information on ethical considerations for this model, please see the Model Card++ Bias, and Privacy Subcards.

Please report model quality, risk, security vulnerabilities or NVIDIA AI Concerns here.

Get Help

NVIDIA Developer Community Forum

Get access to community knowledge base articles and support cases (https://forums.developer.nvidia.com/)

Enterprise Support

Get access to knowledge base articles and support cases or submit a ticket.

Citation

@misc{nvidia_nemotron_3_ultra_2026,\n  title  = {Nemotron 3 Ultra: Open, Efficient Mixture-of-Experts Hybrid Mamba-Transformer Model for Agentic Reasoning},\n  author = {{NVIDIA}},\n  year   = {2026},\n  url    = {https://research.nvidia.com/labs/nemotron/files/NVIDIA-Nemotron-3-Ultra-Technical-Report.pdf},\n  note   = {White Paper}\n}\n

","resourceVersions",[191,192,193,194,195,196,197,198,199,200,201,202,203],"resourceVersionsTotal",12,"hasMoreVersions","versionMetadata",{"_165":21,"_166":167,"_48":168,"_169":-5,"_170":171,"_172":173,"_174":78,"_175":16,"_176":84,"_177":-5,"_178":30,"_179":-5,"_180":30,"_181":-5,"_182":30,"_183":-5,"_184":30,"_185":-5,"_186":30,"_187":-5,"_188":30,"_189":190},"versionNotFound","isLoggedIn","latestVersionPath","/orgs/nim/nvidia/models/nemotron-3-ultra-550b-a55b/-","version","imageReference","cliDownloadCommand","ngc registry model download-version nim/nvidia/nemotron-3-ultra-550b-a55b:1c02e59d-gb200-throughput","helmFetchCommand","viewAllVersionsPath","/orgs/nim/nvidia/models/nemotron-3-ultra-550b-a55b/1c02e59d-gb200-throughput/version-history","accessCheckResult",{"_164":84},"hasRelatedCollections","context",{"_160":161,"_162":163},1789019917530,[],[],[153,154,155,156,157,158,159],{"_44":58,"_46":68,"_48":60,"_50":61,"_52":69,"_54":55},{"_44":58,"_46":66,"_48":60,"_50":61,"_52":67,"_54":55},{"_44":58,"_46":64,"_48":60,"_50":61,"_52":65,"_54":55},{"_44":58,"_46":63,"_48":100,"_50":-5,"_52":-5,"_54":55},{"_44":58,"_46":59,"_48":60,"_50":61,"_52":62,"_54":55},{"_44":56,"_46":57,"_48":49,"_50":51,"_52":53,"_54":55},{"_44":45,"_46":47,"_48":49,"_50":51,"_52":53,"_54":55},"orgName","nim","teamName","nvidia","allowed","versionId","description","gb200 tp4 nvfp4 throughput passthrough","UPLOAD_COMPLETE","releaseType","createdDate",["D",1785274044481],"updatedDate",["D",1785274045157],"totalSizeInBytes","totalFileCount","isSigned","malwareScanStatus","isScanned","releaseNotes","hasReleaseNotes","setup","hasSetup","quickStartGuide","hasQuickStartGuide","advanced","hasAdvanced","performance","hasPerformance","policies",[],{"_74":227,"_76":228,"_169":-5,"_176":84,"_178":30,"_189":226,"_180":30},{"_74":225,"_76":225,"_169":-5,"_176":84,"_178":30,"_189":226,"_180":30},{"_74":223,"_76":223,"_169":-5,"_176":84,"_178":30,"_189":224,"_180":30},{"_74":221,"_76":221,"_169":-5,"_176":84,"_178":30,"_189":222,"_180":30},{"_74":219,"_76":219,"_169":-5,"_176":84,"_178":30,"_189":220,"_180":30},{"_74":217,"_76":217,"_169":-5,"_176":84,"_178":30,"_189":218,"_180":30},{"_74":215,"_76":215,"_169":-5,"_176":84,"_178":30,"_189":216,"_180":30},{"_74":213,"_76":213,"_169":-5,"_176":84,"_178":30,"_189":214,"_180":30},{"_74":211,"_76":211,"_169":-5,"_176":84,"_178":30,"_189":212,"_180":30},{"_74":209,"_76":209,"_169":-5,"_176":84,"_178":30,"_189":210,"_180":30},{"_74":21,"_76":21,"_169":-5,"_176":84,"_178":30,"_189":208,"_180":30},{"_74":206,"_76":206,"_169":-5,"_176":84,"_178":30,"_189":207,"_180":30},{"_74":204,"_76":204,"_169":-5,"_176":84,"_178":30,"_189":205,"_180":30},"hf-590d0da",[],"hf-nvfp4-e9744cc",[],[],"1fd101d2-h100-80gb-nvfp4-tp8",[],"1e941457-b200-lora-fit",[],"0a66f3a2-b300-throughput",[],"f83620b1-h200-nvfp4-tp4",[],"0a66f3a2-gb300-throughput",[],"hf-rl-052726-nvfp4-e9744cc",[],"hf-rl-052726-590d0da",[],"1c02e59d-b200-throughput",[],"901ef672-h100-nvl-nvfp4-tp8",[],"-","Latest (901ef672-h100-nvl-nvfp4-tp8)","nim/nvidia/nemotron-3-ultra-550b-a55b","key","nim/nvidia/models/nemotron-3-ultra-550b-a55b","resourceType","models","resourceId","nemotron-3-ultra-550b-a55b","shortDescription","Nemotron-3-Ultra-550B-A55B-BF16 is a frontier-scale large language model LLM trained by NVIDIA designed to deliver strong agentic reasoning and conversational capabilities.","# NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16\n\n

\n\n $\"Chat\"$ \n\n\n $\"Paper\"$ \n\n\n $\"Pre-Training$ \n\n\n $\"Post-Training$ \n\n

\n \n $\"Homepage\"$ \n \n\n $\"Discord\"$ \n \n

\n\n

\n \n $\"License\"$ \n \n

\n\n![](./accuracy_chart.png)\n\n## Model Summary\n\n| | |\n|:---|:---|\n| **Total Parameters** | 550B (55B active) |\n| **Architecture** | LatentMoE - Mamba-2 + MoE + Attention hybrid with Multi-Token Prediction (MTP) |\n| **Context Length** | Up to 1M tokens |\n| **Minimum GPU Requirement** | 8x GB200/B200/GB300/B300, 16x H100, 8x H200 |\n| **Supported Languages** | English, French, Spanish, Italian, German, Japanese, Korean, Hindi, Korean, Brazilian Portuguese, and Chinese |\n| **Best For** | Frontier reasoning, complex agentic workflows, long-context analysis, tool use, multilingual reasoning, high-stakes RAG |\n| **Reasoning Mode** | Configurable on/off via chat template (`enable_thinking=True/False`) |\n| **License** | [OpenMDW License Agreement, version 1.1](https://raw.githubusercontent.com/OpenMDW/OpenMDW/refs/heads/main/1.1/LICENSE.OpenMDW-1.1) |\n| **Release Date** | June 4, 2026 |\n\n\n## Quick Start\n\nFor more details on how to deploy and use the model - see the [Quick Start Guide](#quick-start-guide) below!\n\n> *For running Nemotron 3 Ultra on a smaller footprint, please see: [NVIDIA-Nemotron-3-Ultra-550B-A55B-NVFP4](https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-NVFP4)*\n\n## Model Overview\n\n**Model Developer:** NVIDIA Corporation\n\n**Model Dates:** December 2025 - April 2026\n\n**Data Freshness:**\n\n* The post-training data has a cutoff date of May 2026.\n* The pre-training data has a cutoff date of September 2025.\n\n### What is Nemotron?\n\nNVIDIA Nemotron™ is a family of open models with open weights, training data, and recipes, delivering leading efficiency and accuracy for building specialized AI agents.\n\n## Description\n\n**Nemotron-3-Ultra-550B-A55B-BF16** is a frontier-scale large language model (LLM) trained by NVIDIA, designed to deliver strong agentic, reasoning, and conversational capabilities. It is optimized for the most demanding workloads, including complex multi-step agents, long-context analysis, and high-accuracy reasoning over code, math, and science. Like other models in the family, it responds to user queries and tasks by first generating a reasoning trace and then concluding with a final response. The model's reasoning capabilities can be configured through a flag in the chat template.\n\nThe model employs a hybrid **Latent Mixture-of-Experts (LatentMoE)** architecture, utilizing interleaved Mamba-2 and MoE layers, along with select Attention layers. Like the Super model, the Ultra model incorporates **Multi-Token Prediction (MTP)** layers for faster text generation and improved quality, and it is trained using an **NVFP4** pre-training recipe to maximize compute efficiency. The model has **55B active parameters** and **550B parameters in total**.\n\nThe supported languages include: English, French, Spanish, Italian, German, Japanese, Korean, Hindi, Korean, Brazilian Portuguese, and Chinese.\n\nThis model is ready for commercial and non-commercial use.\n\n## License/Terms of Use\n\n**Governing Download Terms:** Use of this model is governed by the [OpenMDW License Agreement, version 1.1](https://raw.githubusercontent.com/OpenMDW/OpenMDW/refs/heads/main/1.1/LICENSE.OpenMDW-1.1) (OpenMDW-1.1).\n\n### Benchmarks\n\n| Benchmark | N-3-Ultra
550B-A55B | MiniMax-2.7
230B-A10B | GLM-5.1
744B-A40B | Kimi-K2.6
1T-A32B | Qwen-3.5
397B-17B | DS-v4-Pro
1.6T-A49B | DS-v4-Flash
284B-A13B |\n| :--- | :---: | :---: | :---: | :---: | :---: | :---: | :---: |\n| **Agentic** | | | | | | | |\n| Terminal Bench 2.1 | 56.4 | 55.5 | 59.3 | 67.2 | 49.9 | 49.2 | 54.2 |\n| GDPVal | 46.7 | 47.6 | 54.7 | 50.4 | 34.6 | 54.6 | 50.2 |\n| SWE-Bench Verified | 71.9 | 72.2 | 73.8 | 69.5 | 69.9 | 74.0 | 72.4 |\n| SWE-Bench Multilingual | 67.7 | 69.2 | 73.8 | 65.9 | 67.7 | 71.9 | 72.1 |\n| ProfBench (Search) | 56.0 | 52.0 | 46.0 | 56.0 | 53.0 | 59.9 | 57.0 |\n| PinchBench | 90.0 | 77.6 | 81.2 | 90.2 | 86.6 | 88.6 | 91.3 |\n| TauBench V3 | | | | | | | |\n| Airline | 81.5 | 75.3 | 85.0 | 85.8 | 76.5 | 80.8 | 80.8 |\n| Retail | 86.4 | 84.9 | 84.1 | 82.9 | 88.5 | 88.9 | 89.1 |\n| Telecom | 92.9 | 89.6 | 96.9 | 97.8 | 98.0 | 96.3 | 98.3 |\n| Banking | 22.6 | 14.6 | 12.8 | 23.1 | 20.9 | 25.9 | 26.7 |\n| Average | 70.9 | 66.1 | 69.7 | 72.4 | 71.0 | 73.2 | 73.7 |\n| BrowseComp | 44.4 | 54.1 | 59.4 | 61.3 | 40.5 | 59.4 | 46.9 |\n| Vals.ai Financial Agent 1.1 | | | | | | | |\n| without web search | 60.1 | 51.3 | 60.2 | 54.0 | 61.3 | 58.9 | 58.4 |\n| with web search | 53.7 | 50.5 | 60.7 | 58.8 | 59.0 | 62.3 | 60.1 |\n| **Reasoning and Knowledge** | | | | | | | |\n| IOI 2025 | 570.0 | -- | 456.5 | 585.0 | 441.3 | 580.1 | -- |\n| LiveCodeBench (v6) | 89.0 | 77.2 | 85.7 | 90.2 | 79.3 | 92.5 | 90.9 |\n| IMOAnswerBench (no tools) | 88.6 | 68.3 | 86.8 | 91.1 | 83.1 | 93.0 | 91.1 |\n| IMOAnswerBench (with tools) | 92.3 | 75.1 | 91.1 | 93.71 | 84.51 | 85.4 | 89.6 |\n| Apex-Shortlist (no tools) | 74.9 | 28.9 | 71.1 | 77.4 | 61.4 | 85.8 | 82.4 |\n| Apex-Shortlist (with tools) | 84.8 | 51.9 | 79.0 | 73.2 | 60.4 | 86.5 | 82.0 |\n| GPQA (no tools) | 87.0 | 86.6 | 86.1 | 91.0 | 87.1 | 87.8 | 88.5 |\n| SciCode (subtask) | 44.6 | 38.3 | 47.7 | 52.0 | 48.0 | 50.5 | 48.2 |\n| HLE (no tools) | 26.7 | 23.1 | 27.2 | 34.8 | 28.5 | 37.7 | 32.2 |\n| HLE (with tools) | 37.4 | -- | 50.4 | 54.0 | 48.3 | 48.2 | 45.1 |\n| CritPt (no tools) | 3.1 | 0.6 | 3.7 | 9.1 | 2.4 | 14.0 | 10.6 |\n| MMLU-Pro | 86.8 | 81.9 | 85.9 | 88.1 | 88.3 | 87.5 | 86.4 |\n| OmniScience Accuracy | 24.1 | 20.5 | 31.3 | 35.5 | 35.9 | 46.8 | 39.9 |\n| OmniScience Non-Hallucination | 78.7 | 74.4 | 66.8 | 67.1 | 7.4 | 5.7 | 2.8 |\n| **Chat & Instruction Following** | | | | | | | |\n| IFBench (prompt loose) | 81.7 | 74.6 | 76.6 | 73.7 | 78.2 | 79.1 | 82.0 |\n| Multi-Challenge | 63.8 | 42.5 | 63.0 | 63.1 | 63.9 | 64.1 | 63.5 |\n| **Long Context** | | | | | | | |\n| AA-LCR | 65.4 | 69.8 | 66.9 | 70.2 | 68.3 | 67.3 | 62.7 |\n| RULER (1M) | 94.7 | -- | -- | -- | 90.1 | 94.2 | 87.7 |\n| Longbench v2 (≤ 1M) | 61.9 | -- | -- | -- | 68.9 | 62.1 | 57.0 |\n| **Multilingual** | | | | | | | |\n| MMLU-ProX (avg en/de/fr/es/it/ja/zh/hi/pt/ko) | 83.0 | 78.4 | 85.8 | 85.0 | 86.4 | 85.6 | 84.3 |\n| WMT24++ (en→xx) | 83.7 | 82.8 | 84.4 | 84.5 | 86.8 | 85.9 | 85.9 |\n\nAll evaluation results were collected via [Nemo Evaluator SDK](https://github.com/NVIDIA-NeMo/Evaluator). We used three main evaluation harnesses: [Nemo Gym](https://github.com/NVIDIA-NeMo/Gym), [Nemo Skills](https://github.com/NVIDIA-NeMo/Skills), and [Harbor](https://github.com/harbor-framework/harbor) with extended sandboxing support via AWS ECS on Nemo Evaluator. In addition, the evaluations also used dedicated open-source packaged containers for ScaleAI Multi Challenge Multi Turn Instruction Following and KernelBench. For reproducibility purposes, more details on the evaluation settings and pinned containers can be found in the [Nemo Evaluator SDK examples folder](https://github.com/NVIDIA-NeMo/Evaluator/blob/main/examples/nemotron/nemotron-3-ultra) and the [reproducibility tutorial for Nemotron 3 Ultra](https://github.com/NVIDIA-NeMo/Evaluator/blob/main/examples/nemotron/nemotron-3-ultra/reproducibility.md). \n\nThe following benchmarks are not onboarded yet in our open source tools and for these we used either their official open source implementation or otherwise an internal scaffolding that we plan to open source in the future: BrowseComp with Search, Tau Bench 3, ProfBench with Search, PinchBench, Vals.ai, LongBench v2.\n\n### Deployment Geography: Global\n\n### Use Case\n\nNVIDIA-Nemotron-3-Ultra-550B-A55B-BF16 is a frontier-scale general purpose reasoning and chat model intended to be used in English, Code, and supported multilingual contexts. This model is optimized for complex agentic workflows, long-context reasoning, and high-stakes analytical workloads. It is intended to be used by developers designing AI Agent systems, chatbots, RAG systems, and other AI-powered applications. This model is also suitable for complex instruction-following tasks and long-context reasoning over very large documents and codebases.\n\n### Release Date\n\nHugging Face - 06/04/2026 via [Hugging Face](https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16)\n\n## Reference(s)\n\n* [NVIDIA Nemotron 3 model family on Hugging Face](https://huggingface.co/collections/nvidia/nvidia-nemotron-v3)\n* [NVIDIA Nemotron 3 Ultra Technical Report](https://research.nvidia.com/labs/nemotron/files/NVIDIA-Nemotron-3-Ultra-Technical-Report.pdf)\n\n## Model Architecture\n\n- **Architecture Type:** Mamba2-Transformer Hybrid Latent Mixture of Experts (LatentMoE) with Multi-Token Prediction (MTP)\n- **Network Architecture:** Nemotron Hybrid LatentMoE\n- **Number of model parameters:** 550B Total / 55B Active\n\n## Model Design\n\nThe model utilizes the **LatentMoE** architecture, where tokens are projected into a smaller latent dimension for expert routing and computation, improving accuracy per byte. The Ultra model is pre-trained using an NVFP4 recipe — sharing the quantization-aware pre-training approach pioneered in the Nemotron 3 family. The majority of linear layers use NVFP4 for weights, activations, and gradients, while select layers (including latent projections, MTP layers, QKV/attention projections, and embeddings) are maintained in BF16 or MXFP8 for training stability. The model includes **Multi-Token Prediction (MTP)** layers using a shared-weight design across prediction heads. This improves training signal quality, enables faster inference via native speculative decoding, and supports more stable autoregressive drafting at longer draft lengths compared to independently trained offset heads.\n\n## Training Methodology\n\nStage 1: Pre-Training\n\n* [NVIDIA-Nemotron-3-Ultra-550B-A55B-Base-BF16](https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-Base-BF16) model was pre-trained for approximately 20T tokens using crawled and synthetic code, math, science, and general knowledge data. Training leveraged an NVFP4 recipe for efficiency. All datasets are disclosed in the [Training and Evaluation Datasets](#training-and-evaluation-datasets) section of this document. Major portions of the pre-training corpus are released in the [Nemotron-Pre-Training-Datasets](https://huggingface.co/collections/nvidia/nemotron-pre-training-datasets) collection.\n* Software used for pre-training: [Megatron-LM](https://github.com/NVIDIA/Megatron-LM)\n\nStage 2: Supervised Fine-Tuning\n\n* The model was further fine-tuned on synthetic code, math, science, tool calling, instruction following, structured outputs, and general knowledge data. This stage incorporated data designed to support long-range retrieval and multi-document aggregation. All datasets are disclosed in the [Training and Evaluation Datasets](#training-and-evaluation-datasets) section of this document. Major portions of the fine-tuning corpus are released in the [Nemotron-Post-Training-v3](https://huggingface.co/collections/nvidia/nemotron-post-training-v3) collection. [Data Designer](https://github.com/NVIDIA-NeMo/DataDesigner) is one of the libraries used to prepare these corpora.\n\nStage 3: Reinforcement Learning\n\n* The model underwent multi-environment reinforcement learning using asynchronous GRPO (Group Relative Policy Optimization) across math, code, science, instruction following, multi-step tool use, multi-turn conversations, and structured output environments. It utilized an asynchronous RL architecture that fully decouples training from inference across separate GPU devices, leveraging in-flight weight updates and MTP to accelerate rollout generation. Conversational quality was further refined through RLHF. All datasets are disclosed in the [Training and Evaluation Datasets](#training-and-evaluation-datasets) section of this document. The RL environments and datasets are released as part of [NeMo Gym](https://github.com/NVIDIA-NeMo/Gym).\n* Software used for reinforcement learning: [NeMo RL](https://github.com/NVIDIA-NeMo/RL), [NeMo Gym](https://github.com/NVIDIA-NeMo/Gym)\n\nStage 4: Multi-Domain On-Policy Distillation (MOPD)\n\n* The model underwent **Multi-Domain On-Policy Distillation (MOPD)** to improve reasoning across many task types while staying efficient. This technique uses strong teacher models to guide training on the model's own generated attempts (on-policy rollouts), helping recover accuracy and improve performance across coding, math, instruction following, tool use, and agentic workflows. By distilling teacher signal onto the student's own trajectories rather than offline traces, MOPD better aligns the student's behavior with what it would actually produce at inference time, yielding stronger gains than purely off-policy distillation.\n\nNVIDIA-Nemotron-3-Ultra-550B-A55B-BF16 model is a result of the above work.\n\nThe end-to-end training recipe is available in the [NVIDIA Nemotron Developer Repository](https://github.com/NVIDIA-NeMo/Nemotron). Evaluation results can be replicated using the [NeMo Evaluator SDK](https://github.com/NVIDIA-NeMo/Evaluator). [Data Designer](https://github.com/NVIDIA-NeMo/DataDesigner) is one of the libraries used to prepare the pre and post training datasets. More details on the datasets and synthetic data generation methods can be found in the technical report [NVIDIA Nemotron 3 Ultra Technical Report](https://research.nvidia.com/labs/nemotron/files/NVIDIA-Nemotron-3-Ultra-Technical-Report.pdf).\n\n## Input\n\n- **Input Type(s):** Text\n- **Input Format(s):** String\n- **Input Parameters:** One-Dimensional (1D): Sequences\n- **Other Properties Related to Input:** Maximum context length up to 1M tokens. Supported languages include: English, French, Spanish, Italian, German, Japanese, Korean, Hindi, Korean, Brazilian Portuguese, and Chinese\n\n## Output\n\n- **Output Type(s):** Text\n- **Output Format:** String\n- **Output Parameters:** One-Dimensional (1D): Sequences\n- **Other Properties Related to Output:** Maximum context length up to 1M tokens\n\nOur AI models are designed and optimized to run on NVIDIA GPU-accelerated systems. By leveraging NVIDIA's hardware (e.g. GPU cores) and software frameworks (e.g., CUDA libraries), the model achieves faster training and inference times compared to CPU-only solutions.\n\n## Software Integration\n\n- Runtime Engine(s): NeMo 26.04.01\n- Supported Hardware Microarchitecture Compatibility: NVIDIA Ampere - A100; NVIDIA Blackwell; NVIDIA Hopper - H100-80GB\n- Operating System(s): Linux\n\nThe integration of foundation and fine-tuned models into AI systems requires additional testing using use-case-specific data to ensure safe and effective deployment. Following the V-model methodology, iterative testing and validation at both unit and system levels are essential to mitigate risks, meet technical and functional requirements, and ensure compliance with safety and ethical standards before deployment.\n\n## Model Version(s)\n\n* v1.0 - GA\n\n\nThe Ultra BF16 checkpoint is a frontier-scale model. The minimum recommended hardware is:\n\n* **Single-node:** 8× B200 (≈1.5 TB aggregate HBM — fits BF16 weights plus KV cache with headroom) \n* **Multi-node:** ≥8 GPUs across H100 / H200 / GB200 / GB300, orchestrated with Ray v2\n\nAll deployment snippets below default to **port 8000**, with **chunked prefill** and **MTP (5 speculative tokens)** enabled.\n\n---\n\n#### **Multi-Node Setup with Ray (recommended for multi-node BF16)**\n\nThe recommended multi-processing backend for multi-node BF16 deployments is Ray v2. Below is a template for launching a Ray cluster:\n\n```shell\n# Set the IP for the head node in RAY_HEAD_IP\nexport RAY_HEAD_IP=\nexport RAY_PORT=6379\nexport RAY_ADDRESS=${RAY_HEAD_IP}:${RAY_PORT}\n\n# Start Ray head node (vLLM/SGLang will run on this node)\nray start --head --node-ip-address=${RAY_HEAD_IP} --port=${RAY_PORT}\n\n# Start Ray worker node(s)\nray start --address=${RAY_HEAD_IP}:${RAY_PORT} --block\n\n# Verify Ray cluster is ready\nray status --address=${RAY_HEAD_IP}:${RAY_PORT}\n```\n\n`ray[cgraph]` is required: `uv pip install \"ray[cgraph]\"`\n\n---\n\n#### **vLLM**\n\nRecommended container: `vllm/vllm-openai:v0.22.0`.\n\nFor more detailed information, please see [this cookbook](https://github.com/NVIDIA-NeMo/Nemotron/blob/main/usage-cookbook/Nemotron-3-Ultra/vllm_cookbook.ipynb).\n\n```shell\nexport MODEL_CKPT=PATH/TO/MODEL/CHECKPOINT\n```\n\n**8× B200 single-node deployment:**\n\n```shell\ndocker run -d --name nemotron-ultra-vllm \\\n --gpus all \\\n --ipc=host \\\n --network=host \\\n --shm-size=16g \\\n --ulimit memlock=-1 \\\n --ulimit stack=67108864 \\\n -v $MODEL_CKPT:/model:ro \\\n -e VLLM_WORKER_MULTIPROC_METHOD=spawn \\\n -e SAFETENSORS_FAST_GPU=1 \\\n -e NVIDIA_TF32_OVERRIDE=1 \\\n -e VLLM_LOGGING_LEVEL=INFO \\\n vllm/vllm-openai:v0.22.0 \\\n /model \\\n --host 0.0.0.0 \\\n --port 8000 \\\n --served-model-name nvidia/nemotron-3-ultra \\\n --trust-remote-code \\\n --tensor-parallel-size 8 \\\n --enable-expert-parallel \\\n --dtype bfloat16 \\\n --max-model-len 262144 \\\n --gpu-memory-utilization 0.90 \\\n --max-num-seqs 16 \\\n --max-num-batched-tokens 32768 \\\n --enable-chunked-prefill \\\n --enable-prefix-caching \\\n --reasoning-parser nemotron_3 \\\n --enable-auto-tool-choice \\\n --tool-call-parser qwen3_coder \\\n --mamba-ssm-cache-dtype float16 \\ \n --mamba-backend flashinfer \\ \n --enable-mamba-cache-stochastic-rounding \\ \n --mamba-cache-philox-rounds 5 \\\n --speculative-config '{\"method\": \"nemotron_h_mtp\", \"num_speculative_tokens\": 5}' \\\n --model-loader-extra-config '{\"enable_multithread_load\": true, \"num_threads\": 96}'\n```\n\n**Multi-node deployment (e.g. 2× 4×GB300 with Ray):**\n\nAfter launching the Ray head and worker per the multi-node setup above:\n\n```shell\n# Run on Ray head node\nvllm serve $MODEL_CKPT \\\n --host 0.0.0.0 \\\n --port 8000 \\\n --served-model-name nvidia/nemotron-3-ultra \\\n --tensor-parallel-size 8 \\\n --distributed-executor-backend ray \\\n --trust-remote-code \\\n --dtype bfloat16 \\\n --gpu-memory-utilization 0.90 \\\n --max-model-len 262144 \\\n --max-num-seqs 256 \\\n --max-num-batched-tokens 32768 \\\n --enable-chunked-prefill \\\n --enable-prefix-caching \\\n --reasoning-parser nemotron_3 \\\n --mamba-ssm-cache-dtype float16 \\ \n --mamba-backend flashinfer \\ \n --enable-mamba-cache-stochastic-rounding \\ \n --mamba-cache-philox-rounds 5 \\\n --enable-auto-tool-choice \\\n --tool-call-parser qwen3_coder \\\n --speculative-config '{\"method\": \"nemotron_h_mtp\", \"num_speculative_tokens\": 5}' \\\n --kv-cache-dtype fp8 \\\n --model-loader-extra-config '{\"enable_multithread_load\": true, \"num_threads\": 96}' \\\n --compilation-config '{\"pass_config\": {\"fuse_allreduce_rms\": false}}' \\\n --distributed-timeout-seconds 3600\n```\n\nContext length defaults to 256k above. To use up to 1M, set `VLLM_ALLOW_LONG_MAX_MODEL_LEN=1` and `--max-model-len 1048576`.\n\n**Useful environment variables:** `VLLM_FLASHINFER_ALLREDUCE_BACKEND=trtllm`, `VLLM_FLASHINFER_MOE_BACKEND=latency` (TRTLLM-Gen) or `VLLM_FLASHINFER_MOE_BACKEND=throughput` (CUTLASS).\n\n---\n\n#### **SGLang**\n\nContainer (tested on 8× B200):\n\n```shell\ndocker pull lmsysorg/sglang:v0.5.11\n```\n\nFor more detailed information, please see [this cookbook](https://github.com/NVIDIA-NeMo/Nemotron/blob/main/usage-cookbook/Nemotron-3-Ultra/sglang_cookbook.ipynb).\n\n**8× B200 single-node deployment (BF16, chunked prefill \\+ MTP on by default):**\n\n```shell\ndocker run -d --name nemotron-ultra-sglang \\\n --gpus all \\\n --cap-add SYS_NICE \\\n --ipc=host \\\n --network=host \\\n --shm-size=16g \\\n --ulimit memlock=-1 \\\n --ulimit stack=67108864 \\\n -v $MODEL_CKPT:/model:ro \\\n -e SAFETENSORS_FAST_GPU=1 \\\n -e NVIDIA_TF32_OVERRIDE=1 \\\n -e SGLANG_DISABLE_DEEP_GEMM=1 \\\n lmsysorg/sglang:v0.5.11 \\\n python3 -m sglang.launch_server \\\n --model-path /model \\\n --host 0.0.0.0 \\\n --port 8000 \\\n --served-model-name nvidia/nemotron-3-ultra \\\n --tp-size 8 \\\n --ep-size 8 \\\n --context-length 262144 \\\n --mem-fraction-static 0.85 \\\n --chunked-prefill-size 32768 \\\n --fp8-gemm-backend triton \\\n --moe-runner-backend triton \\\n --mamba-scheduler-strategy no_buffer \\\n --disable-piecewise-cuda-graph \\\n --reasoning-parser nemotron_3 \\\n --tool-call-parser qwen3_coder \\\n --speculative-algorithm EAGLE \\\n --speculative-num-steps 5 \\\n --speculative-eagle-topk 1 \\\n --speculative-num-draft-tokens 5 \\\n --trust-remote-code \\\n --log-level info\n```\n\nContext length defaults to 256k above. To use up to 1M, set `SGLANG_ALLOW_OVERWRITE_LONGER_CONTEXT_LEN=1` and `--context-length 1048576`.\n\n**Tool calls \\+ reasoning parsing:** when calling the chat completions endpoint with `tools`, you **must** set `\"chat_template_kwargs\": {\"enable_thinking\": true, \"force_nonempty_content\": true}` in the request body to parse both reasoning and tool calls correctly.\n\n---\n\n#### **TensorRT-LLM**\n\n**Important Note:** Current support for Nemotron 3 Ultra is limited to NVIDIA Blackwell architecture (including B200/B300 and GB200/GB300). While support for NVIDIA Hopper systems is planned, it is not currently available.\n\nContainer:\n\n```shell\ndocker pull nvcr.io/nvidia/tensorrt-llm/release:1.3.0rc17\n```\n\nFor more detailed information, please see [this cookbook](https://github.com/NVIDIA-NeMo/Nemotron/blob/main/usage-cookbook/Nemotron-3-Ultra/trtllm_cookbook.ipynb).\n\n**8xB200:**\n\n```shell\ncat > ./extra-llm-api-config.yml << EOF\nbackend: pytorch\ntrust_remote_code: true\ntensor_parallel_size: 8\npipeline_parallel_size: 1\ncontext_parallel_size: 1\ngpus_per_node: 8\nmoe_expert_parallel_size: 1\ndisable_overlap_scheduler: false\n\ncuda_graph_config:\n enable_padding: true\n max_batch_size: 256\n\nenable_chunked_prefill: true\nenable_attention_dp: false\nmax_batch_size: 256\nmax_seq_len: null\nmax_num_tokens: 32768\nnum_postprocess_workers: 4\n\nkv_cache_config:\n enable_block_reuse: false\n max_tokens: null\n max_attention_window: null\n sink_token_length: null\n free_gpu_memory_fraction: 0.75\n host_cache_size: null\n cross_kv_cache_fraction: null\n secondary_offload_min_priority: null\n event_buffer_max_size: 0\n attention_dp_events_gather_period_ms: 5\n enable_partial_reuse: true\n copy_on_partial_reuse: true\n use_uvm: false\n max_gpu_total_bytes: 0\n iteration_stats_interval: 1\n dtype: fp8\n tokens_per_block: 32\n mamba_state_cache_interval: 256\n use_kv_cache_manager_v2: false\n max_util_for_resume: 0.95\n\nmoe_config:\n backend: TRTLLM\n max_num_tokens: null\n load_balancer: null\n disable_finalize_fusion: false\n use_low_precision_moe_combine: false\nEOF\n\n\nTLLM_ALLOW_LONG_MAX_MODEL_LEN=1 trtllm-serve \\\n \\\n--max_batch_size 256 \\\n--tp_size 8 --ep_size 1 \\\n--max_num_tokens 32768 \\\n--trust_remote_code \\\n--reasoning_parser nano-v3 \\\n--tool_parser qwen3_coder \\\n--chat_template $MODEL_DIR/chat_template.jinja \\\n--extra_llm_api_options extra-llm-api-config.yml\n```\n\n**Long-context configuration:**\n\nFor long-context benchmarking, set `TLLM_ALLOW_LONG_MAX_MODEL_LEN=1` as an environment variable and add `--max_seq_len ` as the desired maximum context length. The MTP `speculative_config` block above carries over unchanged — on rc16, `max_draft_len` is the authoritative field and `num_nextn_predict_layers` is treated as deprecated.\n\n### **API Client**\n\nThe examples below use the OpenAI-compatible client and work with any of the serving backends above.\n\nNOTE: For coding agents add the following to the API call \\- `extra_body={\"chat_template_kwargs\": {\"force_nonempty_content\": True}}`\n\n```py\nfrom openai import OpenAI\nclient = OpenAI(base_url=\"http://localhost:8000/v1\", api_key=\"EMPTY\")\nMODEL = \"nvidia/nemotron-3-ultra\"\n```\n\n**Reasoning ON (default)**\n\n```py\nresponse = client.chat.completions.create(\n model=MODEL,\n messages=[{\"role\": \"user\", \"content\": \"Write a haiku about GPUs\"}],\n max_tokens=16000,\n temperature=1.0,\n top_p=0.95,\n extra_body={\"chat_template_kwargs\": {\"enable_thinking\": True}}\n)\nprint(response.choices[0].message.content)\n```\n\n**Reasoning OFF**\n\n```py\nresponse = client.chat.completions.create(\n model=MODEL,\n messages=[{\"role\": \"user\", \"content\": \"What is the capital of Japan?\"}],\n max_tokens=16000,\n temperature=1.0,\n top_p=0.95,\n extra_body={\"chat_template_kwargs\": {\"enable_thinking\": False}}\n)\nprint(response.choices[0].message.content)\n```\n\n**Medium-effort reasoning**\n\nUses significantly fewer reasoning tokens than full thinking mode. Recommended as a starting point before tuning explicit token budgets.\n\n```py\nresponse = client.chat.completions.create(\n model=MODEL,\n messages=[{\"role\": \"user\", \"content\": \"What is the capital of Japan?\"}],\n max_tokens=16000,\n temperature=1.0,\n top_p=0.95,\n extra_body={\"chat_template_kwargs\": {\"enable_thinking\": True, \"medium_effort\": True}}\n)\nprint(response.choices[0].message.content)\n```\n\n**Tool calling with reasoning (SGLang requires explicit chat template kwargs)**\n\n```py\nresponse = client.chat.completions.create(\n model=MODEL,\n messages=[{\"role\": \"user\", \"content\": \"What's the weather in New York?\"}],\n tools=[{\n \"type\": \"function\",\n \"function\": {\n \"name\": \"get_weather\",\n \"description\": \"Get the current weather for a city.\",\n \"parameters\": {\n \"type\": \"object\",\n \"properties\": {\n \"city\": {\"type\": \"string\"},\n \"unit\": {\"type\": \"string\", \"enum\": [\"celsius\", \"fahrenheit\"]}\n },\n \"required\": [\"city\"]\n }\n }\n }],\n tool_choice=\"required\",\n max_tokens=256,\n temperature=1.0,\n top_p=0.95,\n extra_body={\"chat_template_kwargs\": {\"enable_thinking\": True, \"force_nonempty_content\": True}}\n)\n```\n\n### **OpenCode**\n\n[OpenCode](https://opencode.ai/docs) is an AI coding agent that runs in your terminal. It connects to any OpenAI-compatible endpoint, making it compatible with all three serving backends above (vLLM, SGLang, and TensorRT-LLM).\n\nCreate or update your `~/.config/opencode/opencode.json`:\n\n```json\n{\n \"$schema\": \"https://opencode.ai/config.json\",\n \"model\": \"local/nvidia-nemotron-3-ultra\",\n \"provider\": {\n \"local\": {\n \"npm\": \"@ai-sdk/openai-compatible\",\n \"name\": \"local_backend\",\n \"options\": {\n \"baseURL\": \"http://localhost:8000/v1\",\n \"apiKey\": \"EMPTY\"\n },\n \"models\": {\n \"nvidia-nemotron-3-ultra\": {\n \"name\": \"nvidia/nemotron-3-ultra\",\n \"limit\": {\n \"context\": 1000000,\n \"output\": 32768\n }\n }\n }\n }\n },\n \"agent\": {\n \"build\": {\n \"temperature\": 1.0,\n \"top_p\": 0.95,\n \"max_tokens\": 32000\n },\n \"plan\": {\n \"temperature\": 1.0,\n \"top_p\": 0.95,\n \"max_tokens\": 32000\n }\n }\n}\n```\n\nAll backends above default to port `8000`, so the `baseURL` works as-is for vLLM, SGLang, and TensorRT-LLM.\n\nTo learn more about other supported agent scaffolds \\- check out [this resource](https://github.com/NVIDIA-NeMo/Nemotron/tree/main/usage-cookbook/Nemotron-3-Ultra/OpenScaffoldingResources)\n\n

Advanced: Budget-Controlled Reasoning

\n\nSet a hard token ceiling on the reasoning trace using `reasoning_budget`. The model will attempt to close the trace at the next newline before the budget is hit; if none is found within 500 tokens it closes abruptly at `reasoning_budget + 500`.\n\n```py\nfrom typing import Any, Dict, List\nimport openai\nfrom transformers import AutoTokenizer\n\n\nclass ThinkingBudgetClient:\n def __init__(self, base_url: str, api_key: str, tokenizer_name_or_path: str):\n self.tokenizer = AutoTokenizer.from_pretrained(tokenizer_name_or_path)\n self.client = openai.OpenAI(base_url=base_url, api_key=api_key)\n\n def chat_completion(\n self,\n model: str,\n messages: List[Dict[str, Any]],\n reasoning_budget: int = 512,\n max_tokens: int = 1024,\n **kwargs,\n ) -> Dict[str, Any]:\n assert max_tokens > reasoning_budget, (\n f\"reasoning_budget must be less than max_tokens. \"\n f\"Got {max_tokens=} and {reasoning_budget=}\"\n )\n\n # Step 1: generate the reasoning trace up to the budget\n response = self.client.chat.completions.create(\n model=model, messages=messages, max_tokens=reasoning_budget, **kwargs\n )\n reasoning_content = response.choices[0].message.content\n if \"\" not in reasoning_content:\n reasoning_content = f\"{reasoning_content}.\\n\\n\\n\"\n\n reasoning_tokens_len = len(\n self.tokenizer.encode(reasoning_content, add_special_tokens=False)\n )\n remaining_tokens = max_tokens - reasoning_tokens_len\n assert remaining_tokens > 0, (\n f\"No tokens remaining for response ({remaining_tokens=}). \"\n \"Increase max_tokens or lower reasoning_budget.\"\n )\n\n # Step 2: continue from the closed reasoning trace\n messages.append({\"role\": \"assistant\", \"content\": reasoning_content})\n prompt = self.tokenizer.apply_chat_template(\n messages, tokenize=False, continue_final_message=True\n )\n response = self.client.completions.create(\n model=model, prompt=prompt, max_tokens=remaining_tokens, **kwargs\n )\n\n return {\n \"reasoning_content\": reasoning_content.strip().strip(\"\").strip(),\n \"content\": response.choices[0].text,\n \"finish_reason\": response.choices[0].finish_reason,\n }\n```\n\n**Example usage** (32-token reasoning budget):\n\n```py\nclient = ThinkingBudgetClient(\n base_url=\"http://localhost:8000/v1\",\n api_key=\"EMPTY\",\n tokenizer_name_or_path=\"nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16\",\n)\n\nresult = client.chat_completion(\n model=\"nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16\",\n messages=[\n {\"role\": \"system\", \"content\": \"You are a helpful assistant. /think\"},\n {\"role\": \"user\", \"content\": \"What is 2+2?\"},\n ],\n reasoning_budget=32,\n max_tokens=512,\n temperature=1.0,\n top_p=0.95,\n)\nprint(result)\n```\n\n

\n\n## Training and Evaluation Datasets\n\n# Training\n\n**Data Modality:** Text \n**The total size:** 53.8 TiB (14.8 trillion tokens) \n**Total number of datasets:** 226 \n**Dataset partition:** *Training [100%], testing [0%], validation [0%]* \n**Time period for training data collection:** 2013 to 2026 \n**Time period for testing data collection:** 2013 to 2026 \n**Time period for validation data collection:** 2013 to 2026 \n**Data Collection Method by dataset:** Hybrid: Automated, Human, Synthetic \n**Labeling Method by dataset:** Hybrid: Automated, Human, Synthetic \n\nNVIDIA-Nemotron-3-Ultra-550B-A55B-BF16 is pre-trained on a large corpus of high-quality curated and synthetically-generated data. It is trained in the English language, as well as 11 other languages and 43 programming languages. Our sources cover a variety of document types such as: webpages, dialogue, articles, and other written materials. The corpus spans domains including legal, math, science, finance, and more. We also include a small portion of question-answering, and alignment style data to improve model accuracy. The model was pre-trained for approximately 20 trillion tokens.\n\nThe post-training corpus for NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16 consists of high-quality curated and synthetically-generated data. Primary languages used for post-training include English, French, Spanish, Italian, German, Japanese, Korean, Hindi, Korean, Brazilian Portuguese, and Chinese.\n\nThese datasets, such as FinePDFs, EssentialWeb, HotpotQA, SQuAD, and HelpSteer3, do not collectively or exhaustively represent all demographic groups (and proportionally therein). For instance, these datasets do not contain explicit mentions of demographic classes such as age, gender, or ethnicity in 64-99% of samples, depending on the source. In the subset where such terms are present, document-based datasets (FinePDFs and EssentialWeb) contain representational skews, such as references to \"male\" outnumbering those to \"female\", and mentions of \"White\" as the most frequent among ethnic identifiers (comprising 43-44% of ethnicity mentions). To mitigate these imbalances, we recommend considering evaluation techniques such as bias audits, fine-tuning with demographically balanced datasets, and mitigation strategies like counterfactual data augmentation to align with the desired model behavior. This evaluation used a 3,000-sample subset per dataset, identified as the optimal threshold for maximizing embedder accuracy.\n\nDuring post-training, we generate synthetic data by distilling trajectories, solutions, and translations from strong teacher models and agent systems, often grounded in real tasks or documents and aggressively filtered for quality. For math, code, and science, we start from curated problem sets and use open source permissive models such as GPT-OSS-120B to produce step-by-step reasoning traces, candidate solutions, best-of-n selection traces, and verified CUDA kernels. For long-context and science, we build synthetic QA and reasoning data by retrieving passages from long documents, generating MCQ/OpenQA questions and answers, and paraphrasing them into multiple prompt/response formats to ensure diversity. Across all pipelines we stack automated verification—compilers, numerical checks, language identification—to ensure our data is high quality.\n\nFor all domains, we apply a unified data filtering pipeline to ensure that only high-quality, license-compliant, and verifiable samples are used for post-training. We first discard malformed examples using structural checks (e.g., missing tool definitions when tool calls are present). We then aggressively filter reasoning traces exhibiting pathological repetition, such as repeated n-grams within a sliding window or across the entire trajectory, which we found to be a strong indicator of malformed or low-quality reasoning. Finally, based on internal audits of synthetically generated datasets, we observed that some teacher models occasionally produce reasoning traces and final responses that implicitly align with specific political entities or promote nationalistic narratives. To mitigate this, we apply targeted keyword- and regex-based filters and remove all trajectories matching such behavior.\n\nAlongside the model, we release our final pre-training and post-training data, as outlined in this section. For ease of analysis, there is a sample set that is ungated. For all remaining code, math and multilingual data, gating and approval is required, and the dataset is permissively licensed for model training purposes.\n\nMore details on the datasets and synthetic data generation methods can be found in the technical report _[**_NVIDIA Nemotron 3 Ultra_**](https://research.nvidia.com/labs/nemotron/files/NVIDIA-Nemotron-3-Ultra-Technical-Report.pdf)_.\n\n

For Detailed Dataset Information: Click here!

\n\n#### **Base Pre-Training Corpus (Nemotron 3 Foundation)**\n\nThe foundation of the model is trained on the **Nemotron-3-Ultra** corpus, comprising the following datasets from the [Nemotron Pretraining Datasets collection](https://huggingface.co/collections/nvidia/nemotron-pre-training-datasets):\n\n| Dataset Collection | Token Counts | Description |\n| :--- | :--- | :--- |\n| **Nemotron-CC-v2** & **v2.1** | 9.1T | A massive collection of English web data filtered from Common Crawl, including 2.5T+ tokens of new organic, translated, and synthetically rephrased content. |\n| **Nemotron-CC-Code-v1** | 427.9B | High-quality code tokens extracted from Common Crawl using the Lynx + LLM pipeline to preserve structure and equations. |\n| **Nemotron-Pretraining-Code-v1** & **v2** & **v3** | 1.7T | Curated GitHub code references with multi-stage filtering, deduplication, and large-scale synthetic code data. |\n| **Nemotron-CC-Math-v1** | 133.3B | High-quality math pre-training dataset preserving LaTeX formatting and mathematical structures. |\n| **Nemotron-Pretraining-Specialized-v1** & **v1.1** & **v1.2** & **Nemotron-Pretraining-SFT-v1** | 660.0B | Synthetic datasets targeting specialized domains such as STEM reasoning and scientific coding. |\n| **Nemotron-Pretraining-Legal-v1** | 4.3B | Synthetic datasets targeting the legal domain. |\n\n### Public Datasets\n\n| Dataset | Collection Period |\n| :---- | :---- |\n| [GSM8K](https://github.com/openai/grade-school-math) | 4/23/2025 |\n| [CC-NEWS](https://commoncrawl.org/blog/news-dataset-available) | 4/23/2025 |\n| [Common Crawl](https://commoncrawl.org/) | 4/23/2025 |\n| [Wikimedia](https://dumps.wikimedia.org/) | 4/23/2025 |\n| [Bespoke-Stratos-17k](https://huggingface.co/datasets/bespokelabs/Bespoke-Stratos-17k) | 4/23/2025 |\n| [tigerbot-kaggle-leetcodesolutions-en-2k](https://huggingface.co/datasets/TigerResearch/tigerbot-kaggle-leetcodesolutions-en-2k) | 4/23/2025 |\n| [glaive-function-calling-v2](https://huggingface.co/datasets/glaiveai/glaive-function-calling-v2) | 4/23/2025 |\n| [APIGen Function-Calling](https://huggingface.co/datasets/Salesforce/xlam-function-calling-60k) | 4/23/2025 |\n| [LMSYS-Chat-1M](https://huggingface.co/datasets/lmsys/lmsys-chat-1m) | 4/23/2025 |\n| [Open Textbook Library \\- CC BY-SA & GNU subset](https://open.umn.edu/opentextbooks/textbooks/) and [OpenStax \\- CC BY-SA subset](https://openstax.org/) | 4/23/2025 |\n| [Advanced Reasoning Benchmark](https://github.com/TheDuckAI/arb), [tigerbot-kaggle-leetcodesolutions-en-2k](https://huggingface.co/datasets/TigerResearch/tigerbot-kaggle-leetcodesolutions-en-2k), [PRM800K](https://github.com/openai/prm800k), and [SciBench](https://github.com/mandyyyyii/scibench) | 4/23/2025 |\n| [FineWeb-2](https://huggingface.co/datasets/HuggingFaceFW/fineweb-2) | 4/23/2025 |\n| [Court Listener](https://www.courtlistener.com/help/api/bulk-data/) | Legacy Download |\n| [peS2o](https://huggingface.co/datasets/allenai/peS2o) | Legacy Download |\n| [OpenWebMath](https://huggingface.co/datasets/open-web-math/open-web-math) | Legacy Download |\n| [BioRxiv](https://www.biorxiv.org/tdm) | Legacy Download |\n| [PMC Open Access Subset](https://pmc.ncbi.nlm.nih.gov/tools/openftlist/) | Legacy Download |\n| [OpenWebText2](https://openwebtext2.readthedocs.io/en/latest/) | Legacy Download |\n| [Stack Exchange Data Dump](https://archive.org/details/stackexchange) | Legacy Download |\n| [PubMed Abstracts](https://github.com/thoppe/The-Pile-PubMed) | Legacy Download |\n| [NIH RePORTER ExPORTER Data Dictionary](https://report.nih.gov/exporter-data-dictionary) | Legacy Download |\n| [arXiv](https://info.arxiv.org/help/bulk_data/index.html) | Legacy Download |\n| [BigScience Workshop Datasets](https://github.com/bigscience-workshop/bigscience/tree/master/train/tr11-176B-ml#datasets) | Legacy Download |\n| [Pushshift Reddit data access](https://api.pushshift.io/guide) | Legacy Download |\n| [SEC's Electronic Data Gathering, Analysis, and Retrieval (EDGAR)](https://www.sec.gov/search-filings) | Legacy Download |\n| [Advanced Mathematical Problem Solving](https://github.com/hendrycks/math?tab=readme-ov-file) | Legacy Download |\n| [MathPile](https://github.com/GAIR-NLP/MathPile/) | Legacy Download |\n| [NuminaMath CoT](https://huggingface.co/datasets/AI-MO/NuminaMath-CoT) | Legacy Download |\n| [PMC Article](https://pmc.ncbi.nlm.nih.gov/tools/textmining/) | Legacy Download |\n| [FLAN](https://github.com/google-research/FLAN) | Legacy Download |\n| [Advanced Reasoning Benchmark](https://github.com/TheDuckAI/arb) | Legacy Download |\n| [SciBench](https://github.com/mandyyyyii/scibench) | Legacy Download |\n| [WikiTableQuestions](https://huggingface.co/datasets/stanfordnlp/wikitablequestions) | Legacy Download |\n| [FinQA](https://finqasite.github.io/) | Legacy Download |\n| [Riddles](https://github.com/crawsome/riddles) | Legacy Download |\n| [Problems in Elementary Mathematics for Home Study](https://archive.org/details/AntonovVygodskyNikitinSankinProblemsInElementaryMathematicsForHomeStudyMir1982) | Legacy Download |\n| [MedMCQA](https://huggingface.co/datasets/openlifescienceai/medmcqa) | Legacy Download |\n| [Cosmos QA](https://huggingface.co/datasets/allenai/cosmos_qa) | Legacy Download |\n| [MCTest](https://huggingface.co/datasets/sagnikrayc/mctest) | Legacy Download |\n| [AI2's Reasoning Challenge](https://huggingface.co/datasets/allenai/ai2_arc) | Legacy Download |\n| [OpenBookQA](https://github.com/allenai/OpenBookQA) | Legacy Download |\n| [MMLU Auxiliary Train](https://huggingface.co/datasets/cais/mmlu/viewer/all/auxiliary_train) | Legacy Download |\n| [social-chemestry-101](https://huggingface.co/datasets/tasksource/social-chemestry-101) | Legacy Download |\n| [Moral Stories](https://huggingface.co/datasets/demelin/moral_stories) | Legacy Download |\n| [The Common Pile v0.1](https://huggingface.co/common-pile) | Legacy Download |\n| [FineMath](https://huggingface.co/datasets/HuggingFaceTB/finemath) | Legacy Download |\n| [MegaMath](https://huggingface.co/datasets/LLM360/MegaMath) | Legacy Download |\n| [MegaMath](https://huggingface.co/datasets/LLM360/MegaMath) | Legacy Download |\n| [MultiverseMathHard](https://huggingface.co/datasets/Nexusflow/MultiverseMathHard) | 10/2/2025 |\n| [News Commentary](https://opus.nlpl.eu/datasets/News-Commentary) | 10/2/2025 |\n| [Essential-Web](https://huggingface.co/datasets/EssentialAI/essential-web-v1.0) | 10/2/2025 |\n| [finepdfs](https://huggingface.co/datasets/HuggingFaceFW/finepdfs) | 10/2/2025 |\n| [HotpotQA](https://huggingface.co/datasets/hotpotqa/hotpot_qa) | 10/2/2025 |\n| [SQuAD2.0](https://rajpurkar.github.io/SQuAD-explorer/) | 10/2/2025 |\n| [NLTK Words Lists](https://www.nltk.org/nltk_data/) | 10/2/2025 |\n\n### **Crawled and Scraped from Online Sources by NVIDIA**\n\nThe English Common Crawl data was downloaded from the Common Crawl Foundation (see their FAQ for details on their crawling) and includes the snapshots CC-MAIN-2013-20 through CC-MAIN-2025-13. The data was subsequently deduplicated and filtered in various ways described in the Nemotron-CC paper. Additionally, we extracted data for fifteen languages from the following three Common Crawl snapshots: CC-MAIN-2024-51, CC-MAIN-2025-08, CC-MAIN-2025-18. The fifteen languages included were Arabic, Chinese, Danish, Dutch, French, German, Italian, Japanese, Korean, Polish, Portuguese, Russian, Spanish, Swedish, and Thai. As we did not have reliable multilingual model-based quality classifiers available, we applied just heuristic filtering instead—similar to what we did for lower quality English data in the Nemotron-CC pipeline, but selectively removing some filters for some languages that did not work well. Deduplication was done in the same way as for Nemotron-CC.\n\nThe GitHub Crawl was collected using the GitHub REST API and the Amazon S3 API. Each crawl was operated in accordance with the rate limits set by its respective source, either GitHub or S3. We collect raw source code and subsequently remove any having a license which does not exist in our permissive-license set (for additional details, refer to the [technical report](https://research.nvidia.com/labs/nemotron/files/NVIDIA-Nemotron-3-Ultra-Technical-Report.pdf)).\n\n| Dataset | Modality | Dataset Size | Collection Period | Collecting Organisation |\n| :---- | :---- | :---- | :---- | :---- |\n| English Common Crawl | Text | 3.36T | 4/8/2025 | NVIDIA Advanced Deep Learning Research |\n| English Common Crawl 1.1 | Text | Not disclosed | 10/2/2025 | NVIDIA Advanced Deep Learning Research |\n| Multilingual Common Crawl | Text | 812.7B | 5/1/2025 | NVIDIA Advanced Deep Learning Research |\n| GitHub Crawl | Text | 747.4B | 4/29/2025 | NVIDIA Advanced Deep Learning Research |\n| GitHub Crawl 1.1 | Text | 172.7B | 9/30/2025 | NVIDIA Advanced Deep Learning Research |\n\n## Private Non-publicly Accessible Datasets of Third Parties\n\n| Dataset | Model(s) used |\n|---------|---------------|\n| Global Regulation | Unknown |\n| TAUS Translation Memory | Unknown |\n| Scale HLE | Unknown |\n| HackerRank Coding | Unknown |\n| RL data for Search | Gemini 3; GPT-5 |\n\n## Private Non-publicly Accessible Datasets by NVIDIA\n\n| Dataset | Model(s) used |\n|---------|---------------|\n| Simple Minesweeper | Undisclosed |\n| Simple Sudoku | Undisclosed |\n| Multitool Typewriter Hard | Undisclosed |\n| Machine Translation of News Commentary and TAUS Translation Memory | Undisclosed |\n| Machine Translation of STEM - | [Qwen2.5-14B-Instruct](https://huggingface.co/Qwen/Qwen2.5-14B-Instruct) |\n| Competitive Coding RL data from Nemotron Cascade | Undisclosed |\n| Long context RL | Undisclosed |\n| Single-step SWE RL for patch generation | Undisclosed |\n| OpenHands SWE | Undisclosed |\n\n## NVIDIA-Sourced Synthetic Datasets (Pre-Training)\n\n| Dataset | Modality | Dataset Size | Seed Dataset | Model(s) used for generation |\n| :---- | :---- | :---- | :---- | :---- |\n| Nemotron-Pretraining-Fact-Seeking | Text | 35.0B | [FineWiki](https://huggingface.co/datasets/HuggingFaceFW/finewiki) | [Qwen3-30B-A3B-Instruct-2507](https://huggingface.co/Qwen/Qwen3-30B-A3B-Instruct-2507) |\n| Nemotron-Pretraining-Legal | Text | 4.3B | CommonPile (caselaw_access_project_filtered); California Code of Regulations; Judicial Ethics Opinions; [GLOBALCIT](https://globalcit.eu/); [CUAD](https://www.atticusprojectai.org/cuad); Nemotron Personas; [ToSDR Terms of Service Corpus](https://www.kaggle.com/datasets/sonu1607/tosdr-terms-of-service-corpus); CodeHima/TOS_Dataset; [ContractNLI](https://stanfordnlp.github.io/contract-nli/); CaseHOLD; Code of Federal Regulations; [Canadian Case Law](https://huggingface.co/datasets/a2aj/canadian-case-law) (subsets that allow commercial use) | [Qwen3-235B-A22B-Thinking-2507](https://huggingface.co/Qwen/Qwen3-235B-A22B-Thinking-2507) |\n| Nemotron-Pretraining-Formal-Logic | Text | 128M | [Nemotron Personas](https://huggingface.co/datasets/nvidia/Nemotron-Personas-USA) | [Qwen3-235B-A22B-Thinking-2507](https://huggingface.co/Qwen/Qwen3-235B-A22B-Thinking-2507) |\n| Nemotron-Pretraining-Economics | Text | 73.4M | - | [Qwen3-235B-A22B-Thinking-2507](https://huggingface.co/Qwen/Qwen3-235B-A22B-Thinking-2507) |\n| Nemotron-Pretraining-Multiple-Choice | Text | 1.6B | [MMLU Auxiliary Train](https://huggingface.co/datasets/cais/mmlu/viewer/all/auxiliary_train) | [DeepSeek-V3](https://huggingface.co/deepseek-ai/DeepSeek-V3); [Qwen3-235B-A22B](https://huggingface.co/Qwen/Qwen3-235B-A22B) |\n| Nemotron-Pretraining-Code-Concepts | Text | 7.3B | - | [gpt-oss-20b](https://huggingface.co/openai/gpt-oss-20b); [gpt-oss-120b](https://huggingface.co/openai/gpt-oss-120b) |\n| Nemotron-Pretraining-Unconditional-Algorithmic | Text | 196.5M | - | [gpt-oss-120b](https://huggingface.co/openai/gpt-oss-120b); [Qwen3-235B-A22B](https://huggingface.co/Qwen/Qwen3-235B-A22B) |\n| More Synthetic Tasks from DeepSeek-V3 and Qwen3-235B-A22B | Text | 1.1B | train splits of acp_bench; ai2_arc; babi; gsm8k; hendrycks_math; IFEval; MedText; mediqa_qa; mlqa; MMLU-Pro; mmlu-pro-plus; MMLU-ProX; nq_open; tinyGSM8k; truthful_qa; truthfulqa-multi; MATH-lighteval; mmlu; awesome-chatgpt-prompts; super_glue | [DeepSeek v3](https://huggingface.co/deepseek-ai/DeepSeek-V3); [Qwen3-235B-A22B](https://huggingface.co/Qwen/Qwen3-235B-A22B) |\n| Synthetic Tasks from DeepSeek-V3 and Qwen3-235B-A22B | Text | 6.7B | train splits of Into the Unknown; AI2 ARC (AI2 Reasoning Challenge); BLiMP (Benchmark of Linguistic Minimal Pairs); CommonSenseQA; GLUE; HeadQA; Hendrycks Ethics; Memo Trap; modus-tollens; NeQA; pattern-matching-suppression; mastermind_24_mcq_random; mastermind_24_mcq_close; quote-repetition; redefine-math; Repetitive Algebra; sig-figs; MMLU-Pro; MC-TACO; MedConceptsQA; MMLU_dataset; OpenbooksQA; PIQA (Physical Interaction Question Answering); SocialIQA; SuperGLUE; tinyAI2_arc; tinyMMLU; tinyWinogrande; TruthfulQA; WebQuestions; Winogrande; GPQA; MBPP | [DeepSeek v3](https://huggingface.co/deepseek-ai/DeepSeek-V3); [Qwen3-235B-A22B](https://huggingface.co/Qwen/Qwen3-235B-A22B) |\n| Synthetic Art of Problem Solving from DeepSeek-R1 | Text | 40B | [Art of Problem Solving](https://artofproblemsolving.com/company); [American Mathematics Competitions 8](https://artofproblemsolving.com/wiki/index.php/AMC_8_Problems_and_Solutions); [American Mathematics Competitions 10](https://artofproblemsolving.com/wiki/index.php/AMC_10_Problems_and_Solutions); | [DeepSeek-R1](https://huggingface.co/deepseek-ai/DeepSeek-R1) |\n| Synthetic Moral Stories and Social Chemistry from Qwen3-235B-A22B-Thinking-2507 and Mixtral-8x22B-v0.1 | Text | 15.2M | [social-chemestry-101](https://huggingface.co/datasets/tasksource/social-chemestry-101); [Moral Stories](https://huggingface.co/datasets/demelin/moral_stories) | [Qwen3-235B-A22B-Thinking-2507](https://huggingface.co/Qwen/Qwen3-235B-A22B-Thinking-2507); [Mixtral-8x22B-v0.1](https://huggingface.co/mistralai/Mixtral-8x22B-v0.1) |\n| Synthetic Moral Stories and Social Chemistry from Mixtral-8x22B-v0.1 | Text | 327M | [social-chemestry-101](https://huggingface.co/datasets/tasksource/social-chemestry-101); [Moral Stories](https://huggingface.co/datasets/demelin/moral_stories) | [Mixtral-8x22B-v0.1](https://huggingface.co/mistralai/Mixtral-8x22B-v0.1) |\n| Synthetic Social Sciences seeded with OpenStax from DeepSeek-V3, Mixtral-8x22B-v0.1, and Qwen2.5-72B | Text | 83.6M | [OpenStax \\- CC BY-SA subset](https://openstax.org/) | [DeepSeek-V3](https://huggingface.co/deepseek-ai/DeepSeek-V3); [Mixtral-8x22B-v0.1](https://huggingface.co/mistralai/Mixtral-8x22B-v0.1); [Qwen2.5-72B](https://huggingface.co/Qwen/Qwen2.5-72B) |\n| Synthetic Health Sciences seeded with OpenStax from DeepSeek-V3, Mixtral-8x22B-v0.1, and Qwen2.5-72B | Text | 9.7M | [OpenStax \\- CC BY-SA subset](https://openstax.org/) | [DeepSeek-V3](https://huggingface.co/deepseek-ai/DeepSeek-V3); [Mixtral-8x22B-v0.1](https://huggingface.co/mistralai/Mixtral-8x22B-v0.1); [Qwen2.5-72B](https://huggingface.co/Qwen/Qwen2.5-72B) |\n| Synthetic STEM seeded with OpenStax, Open Textbook Library, and GSM8K from DeepSeek-R1, DeepSeek-V3, DeepSeek-V3-0324, and Qwen2.5-72B | Text | 175M | [OpenStax \\- CC BY-SA subset](https://openstax.org/); [GSM8K](https://github.com/openai/grade-school-math); [Open Textbook Library \\- CC BY-SA & GNU subset](https://open.umn.edu/opentextbooks/textbooks/) | [DeepSeek-R1](https://huggingface.co/deepseek-ai/DeepSeek-R1), [DeepSeek-V3](https://huggingface.co/deepseek-ai/DeepSeek-V3); [DeepSeek-V3-0324](https://huggingface.co/deepseek-ai/DeepSeek-V3-0324); [Qwen2.5-72B](https://huggingface.co/Qwen/Qwen2.5-72B) |\n| [Nemotron-PrismMath](https://huggingface.co/datasets/nvidia/Nemotron-PrismMath) | Text | 4.6B | [Big-Math-RL-Verified](https://huggingface.co/datasets/SynthLabsAI/Big-Math-RL-Verified); [OpenR1-Math-220k](https://huggingface.co/datasets/open-r1/OpenR1-Math-220k) | [Qwen2.5-0.5B-instruct](https://huggingface.co/Qwen/Qwen2.5-0.5B-Instruct), [Qwen2.5-72B-Instruct](https://huggingface.co/Qwen/Qwen2.5-72B-Instruct); [DeepSeek-R1-Distill-Qwen-32B](https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-32B) |\n| Synthetic Question Answering Data from Papers and Permissible Books from Qwen2.5-72B-Instruct | Text | 350M | [arXiv](https://info.arxiv.org/help/bulk_data/index.html); [National Institutes of Health ExPorter](https://www.nih.gov/); [BioRxiv](https://www.biorxiv.org/tdm); [PMC Article](https://pmc.ncbi.nlm.nih.gov/tools/textmining/); [USPTO Backgrounds](https://data.uspto.gov/apis/transition-guide/bdss#pats); [peS2o](https://huggingface.co/datasets/allenai/peS2o); Global Regulation; [CORE](https://core.ac.uk/documentation/dataset); [PG-19](https://github.com/google-deepmind/pg19); [DOAB CC BY & CC BY-SA subset](https://www.doabooks.org/en); [NDLTD ETD resources](https://ndltd.org/thesis-resources/find-etds/) | [Qwen2.5-72B-Instruct](https://huggingface.co/Qwen/Qwen2.5-72B-Instruct) |\n| Synthetic Rephrased [Math Data from Common Crawl](https://huggingface.co/datasets/nvidia/Nemotron-MIND) from phi-4 | Text | 73B | [Common Crawl](https://commoncrawl.org/latest-crawl) | [phi-4](https://huggingface.co/microsoft/phi-4) |\n| Synthetic Math Data from Common Crawl 4plus | Text | 52.3B | [Common Crawl](https://commoncrawl.org/latest-crawl) | [phi-4](https://huggingface.co/microsoft/phi-4) |\n| Synthetic Math Data from Common Crawl 3 | Text | 80.9B | [Common Crawl](https://commoncrawl.org/latest-crawl) | [phi-4](https://huggingface.co/microsoft/phi-4) |\n| Synthetic AGIEval seeded with AQUA-RAT, LogiQA, and AR-LSAT from DeepSeek-V3 and DeepSeek-V3-0324 | Text | 4.0B | [AQUA-RAT](https://huggingface.co/datasets/deepmind/aqua_rat); [LogiQA](https://huggingface.co/datasets/lucasmccabe/logiqa); [AR-LSAT](https://github.com/zhongwanjun/AR-LSAT) | [DeepSeek-V3](https://huggingface.co/deepseek-ai/DeepSeek-V3); [DeepSeek-V3-0324](https://huggingface.co/deepseek-ai/DeepSeek-V3-0324) |\n| Synthetic AGIEval seeded with AQUA-RAT, LogiQA, and AR-LSAT from Qwen3-30B-A3B | Text | 4.2B | [AQUA-RAT](https://huggingface.co/datasets/deepmind/aqua_rat); [LogiQA](https://huggingface.co/datasets/lucasmccabe/logiqa); [AR-LSAT](https://github.com/zhongwanjun/AR-LSAT) | [Qwen3-30B-A3B](https://huggingface.co/Qwen/Qwen3-30B-A3B) |\n| Synthetic Art of Problem Solving from Qwen2.5-32B-Instruct, Qwen2.5-Math-72B, Qwen2.5-Math-7B, and Qwen2.5-72B-Instruct | Text | Undisclosed | [Art of Problem Solving](https://artofproblemsolving.com/company); [American Mathematics Competitions 8](https://artofproblemsolving.com/wiki/index.php/AMC_8_Problems_and_Solutions); [American Mathematics Competitions 10](https://artofproblemsolving.com/wiki/index.php/AMC_10_Problems_and_Solutions); [GSM8K](https://github.com/openai/grade-school-math); [PRM800K](https://github.com/openai/prm800k) | [Qwen2.5-32B-Instruct](https://huggingface.co/Qwen/Qwen2.5-32B-Instruct); [Qwen2.5-Math-72B](https://huggingface.co/Qwen/Qwen2.5-Math-72B); [Qwen2.5-Math-7B](https://huggingface.co/Qwen/Qwen2.5-Math-7B); [Qwen2.5-72B-Instruct](https://huggingface.co/Qwen/Qwen2.5-72B-Instruct) |\n| Synthetic MMLU Auxiliary Train from DeepSeek-R1 | Text | 0.5B | [MMLU Auxiliary Train](https://huggingface.co/datasets/cais/mmlu/viewer/all/auxiliary_train) | [DeepSeek-R1](https://huggingface.co/deepseek-ai/DeepSeek-R1) |\n| Synthetic Long Context Continued Post-Training Data from Papers and Permissible Books from Qwen2.5-72B-Instruct | Text | Undisclosed | [arXiv](https://info.arxiv.org/help/bulk_data/index.html); [National Institutes of Health ExPorter](https://www.nih.gov/); [BioRxiv](https://www.biorxiv.org/tdm); [PMC Article](https://pmc.ncbi.nlm.nih.gov/tools/textmining/); [USPTO Backgrounds](https://data.uspto.gov/apis/transition-guide/bdss#pats); [peS2o](https://huggingface.co/datasets/allenai/peS2o); Global Regulation; [CORE](https://core.ac.uk/documentation/dataset); [PG-19](https://github.com/google-deepmind/pg19); [DOAB CC BY & CC BY-SA subset](https://www.doabooks.org/en); [NDLTD ETD resources](https://ndltd.org/thesis-resources/find-etds/) | [Qwen2.5-72B-Instruct](https://huggingface.co/Qwen/Qwen2.5-72B-Instruct) |\n| Synthetic Common Crawl from Qwen3-30B-A3B and Mistral-Nemo-12B-Instruct | Text | 415.8B | [Common Crawl](https://commoncrawl.org/) | [Qwen3-30B-A3B](https://huggingface.co/Qwen/Qwen3-30B-A3B); [Mistral-NeMo-12B-Instruct](https://huggingface.co/nvidia/Mistral-NeMo-12B-Instruct) |\n| Synthetic Multilingual Data from Common Crawl from Qwen3-30B-A3B | Text | Undisclosed | [Common Crawl](https://commoncrawl.org/) | [Qwen3-30B-A3B](https://huggingface.co/Qwen/Qwen3-30B-A3B) |\n| Synthetic Multilingual Data from Wikimedia from Qwen3-30B-A3B | Text | Undisclosed | [Wikimedia](https://dumps.wikimedia.org/) | [Qwen3-30B-A3B](https://huggingface.co/Qwen/Qwen3-30B-A3B) |\n| Synthetic Math Data from Wikimedia from Nemotron-4-340B-Instruct | Text | Undisclosed | \\- | [Nemotron-4-340B-Instruct](https://huggingface.co/nvidia/Nemotron-4-340B-Instruct) |\n| Synthetic Common Crawl Code from phi-4 | Text | 427.9B | [Common Crawl](https://commoncrawl.org/latest-crawl) | [phi-4](https://huggingface.co/microsoft/phi-4) |\n| Synthetic Scientific Coding from Qwen3-235B-A22B | Text | 1.2B | [Wikimedia](https://dumps.wikimedia.org/) | [Qwen3-235B-A22B](https://huggingface.co/Qwen/Qwen3-235B-A22B-Instruct-2507) |\n| Tool Calling Data | Text | 26.2B | | [Qwen3-235B-A22B-2507](https://huggingface.co/Qwen/Qwen3-235B-A22B-Instruct-2507); [gpt-oss-120b](https://huggingface.co/openai/gpt-oss-120b) |\n| Synthetic Essential-Web from QwQ-32B | Text | 28.1B | [Essential-Web](https://huggingface.co/datasets/EssentialAI/essential-web-v1.0) | [QwQ-32B](https://huggingface.co/Qwen/QwQ-32B) |\n| Translated Synthetic Crawl | Text | 389.9B | [Common Crawl](https://commoncrawl.org/) | [Qwen3-30B-A3B](https://huggingface.co/Qwen/Qwen3-30B-A3B) |\n| Translated Synthetic Wikipedia | Text | 7.9B | [Wikimedia](https://dumps.wikimedia.org/) | [Qwen3-30B-A3B](https://huggingface.co/Qwen/Qwen3-30B-A3B) |\n| Synthetic Long Context from Qwen3-235B-A22B-Instruct-2507 | Text | Undisclosed | [CORE](https://core.ac.uk/documentation/dataset); [PG-19](https://github.com/google-deepmind/pg19); [DOAB CC BY & CC BY-SA subset](https://www.doabooks.org/en); [NDLTD ETD resources](https://ndltd.org/thesis-resources/find-etds/) | [Qwen3-235B-A22B-Instruct-2507](https://huggingface.co/Qwen/Qwen3-235B-A22B-Instruct-2507) |\n| Synthetic Search STEM OPENQ from DeepSeek-R1-0528 | Text | Undisclosed | - | [DeepSeek-R1-0528](https://huggingface.co/deepseek-ai/DeepSeek-R1-0528) |\n| Synthetic MCQ from Qwen2.5-32B-Instruct and DeepSeek-R1-0528 | Text | Undisclosed | - | [Qwen2.5-32B-Instruct](https://huggingface.co/Qwen/Qwen2.5-32B-Instruct); [DeepSeek-R1-0528](https://huggingface.co/deepseek-ai/DeepSeek-R1-0528) |\n| Synthetic Offline Search MCQA HLE from DeepSeek-R1-0528 | Text | Undisclosed | - | [DeepSeek-R1-0528](https://huggingface.co/deepseek-ai/DeepSeek-R1-0528) |\n| Synthetic Offline Search MCQA GPQA from Qwen3-235B-A22B and DeepSeek-R1-0528 | Text | Undisclosed | - | [Qwen3-235B-A22B](https://huggingface.co/Qwen/Qwen3-235B-A22B); [DeepSeek-R1-0528](https://huggingface.co/deepseek-ai/DeepSeek-R1-0528) |\n| Synthetic Human Preference from QwQ-32B, Qwen3-30B-A3B, Qwen3-235B-A22B, Qwen3-235B-A22B-Instruct-2507, Mistral-Small-3.1-24B-Instruct-2503, Mistral-Small-3.2-24B-Instruct-2506, MiniMax-M1-80k, MiniMax-M1-40k, Kimi-K2-Instruct, DeepSeek-V3-0324, DeepSeek-R1-0528 | Text | Undisclosed | - | [QwQ-32B](https://huggingface.co/Qwen/QwQ-32B); [Qwen3-30B-A3B](https://huggingface.co/Qwen/Qwen3-30B-A3B); [Qwen3-235B-A22B](https://huggingface.co/Qwen/Qwen3-235B-A22B); [Qwen3-235B-A22B-Instruct-2507](https://huggingface.co/Qwen/Qwen3-235B-A22B-Instruct-2507); [Mistral-Small-3.1-24B-Instruct-2503](https://huggingface.co/mistralai/Mistral-Small-3.1-24B-Instruct-2503); [Mistral-Small-3.2-24B-Instruct-2506](https://huggingface.co/mistralai/Mistral-Small-3.2-24B-Instruct-2506); [MiniMax-M1-80k](https://huggingface.co/MiniMaxAI/MiniMax-M1-80k); [MiniMax-M1-40k](https://huggingface.co/MiniMaxAI/MiniMax-M1-40k); [Kimi-K2-Instruct](https://huggingface.co/moonshotai/Kimi-K2-Instruct); [DeepSeek-V3-0324](https://huggingface.co/deepseek-ai/DeepSeek-V3-0324); [DeepSeek-R1-0528](https://huggingface.co/deepseek-ai/DeepSeek-R1-0528) |\n| Synthetic Code from Qwen3-32B | Text | Undisclosed | English Common Crawl; English Common Crawl 1.1 | [Qwen3-32B](https://huggingface.co/Qwen/Qwen3-32B) |\n| Synthetic OpenCodeReasoning from DeepSeek-R1 | Text | Undisclosed | [OpenCodeReasoning](https://huggingface.co/datasets/nvidia/OpenCodeReasoning) | [DeepSeek-R1](https://huggingface.co/deepseek-ai/DeepSeek-R1) |\n| Synthetic LIMO from DeepSeek-R1-0528 | Text | Undisclosed | [LIMO](https://huggingface.co/datasets/GAIR/LIMO) | [DeepSeek-R1-0528](https://huggingface.co/deepseek-ai/DeepSeek-R1-0528) |\n| Synthetic SCP from DeepSeek-R1-0528 | Text | Undisclosed | [SCP-116K](https://huggingface.co/datasets/EricLu/SCP-116K) | [DeepSeek-R1-0528](https://huggingface.co/deepseek-ai/DeepSeek-R1-0528) |\n| Synthetic Stack Exchange from DeepSeek-R1-0528 | Text | Undisclosed | [Stack Exchange](https://archive.org/details/stackexchange) | [DeepSeek-R1-0528](https://huggingface.co/deepseek-ai/DeepSeek-R1-0528) |\n| Synthetic Common Crawl from Qwen3-30B-A3B | Text | Undisclosed | [Common Crawl](https://commoncrawl.org/) | [Qwen3-30B-A3B](https://huggingface.co/Qwen/Qwen3-30B-A3B) |\n| Synthetic Wikipedia from Qwen3-30B-A3B | Text | Undisclosed | [Wikimedia](https://dumps.wikimedia.org/) | [Qwen3-30B-A3B](https://huggingface.co/Qwen/Qwen3-30B-A3B) |\n| Synthetic Essential-Web from Qwen3-30B-A3B and Qwen3-235B-A22B-Thinking-2507 | Text | Undisclosed | [Essential-Web](https://huggingface.co/datasets/EssentialAI/essential-web-v1.0) | [Qwen3-30B-A3B](https://huggingface.co/Qwen/Qwen3-30B-A3B); [Qwen3-235B-A22B-Thinking-2507](https://huggingface.co/Qwen/Qwen3-235B-A22B-Thinking-2507) |\n| Synthetic Textbook Math from Qwen3-30B-A3B, Qwen3-235B-A22B, phi-4 | Text | Undisclosed | [Common Crawl](https://commoncrawl.org/); [FineMath](https://huggingface.co/datasets/HuggingFaceTB/finemath) | [Qwen3-30B-A3B](https://huggingface.co/Qwen/Qwen3-30B-A3B); [Qwen3-235B-A22B](https://huggingface.co/Qwen/Qwen3-235B-A22B); [phi-4](https://huggingface.co/microsoft/phi-4) |\n| Synthetic Math and Code from DeepSeek-R1 and DeepSeek-R1-0528 | Text | Undisclosed | [Magicoder-Evol-Instruct-110K](https://huggingface.co/datasets/ise-uiuc/Magicoder-Evol-Instruct-110K); [opc-sft-stage2](https://huggingface.co/datasets/OpenCoder-LLM/opc-sft-stage2); [TACO](https://huggingface.co/datasets/BAAI/TACO); [OpenCodeReasoning](https://huggingface.co/datasets/nvidia/OpenCodeReasoning); [OpenMathReasoning](https://huggingface.co/datasets/nvidia/OpenMathReasoning); [NuminaMath CoT](https://huggingface.co/datasets/AI-MO/NuminaMath-CoT) | [DeepSeek-R1](https://huggingface.co/deepseek-ai/DeepSeek-R1); [DeepSeek-R1-0528](https://huggingface.co/deepseek-ai/DeepSeek-R1-0528) |\n\n
\n\n## NVIDIA-Sourced Synthetic Datasets (Post-Training)\n| Dataset | Modality | Dataset Size | Seed Dataset | Model(s) used for generation |\n| :---- | :---- | :---- | :---- | :---- |\n| Synthetic Competitive MATH Proofs from DeepSeek-V4-Pro | Text | Undisclosed | [AMC 8 Problems and Solutions, AMC 10 Problems and Solution, and AIME Problems and Solutions] | [deepseek-ai/DeepSeek-V4-Pro] |\n| Synthetic Hermes Agent Reasoning Traces | Text | Undisclosed | [lambda/hermes-agent-reasoning-traces] | [hermes-agent-generator] |\n| Synthetic Competitive Coding from DeepSeek-V4-Pro | Text | Undisclosed | [NVCompetitiveCodingV1] | [deepseek-ai/DeepSeek-V4-Pro] |\n| Synthetic Competitive Science Reasoning from DeepSeek-V4-Pro | Text | Undisclosed | [AMC 8 Problems and Solutions, AMC 10 Problems and Solution, and AIME Problems and Solutions]; [EssentialAI/essential-web-v1.0]; [cdquestions.com]; [Pile-FreeLaw]; [Vedantu]; [askfilo]; [doubtnut]; [ICHO-IPH0 Dataset]; [LIMO dataset (Less is More for Reasoning)]; [AAPT]; [ChemData 700K]; [oMeBench]; [Flavor Analysis and Recognition Transformer]; [ChemCoTBench]; [Llama Nemotron Dataset] | [deepseek-ai/DeepSeek-V4-Pro] |\n| Synthetic Competitive MATH CoT and TIR from Nemotron 5.5 | Text | Undisclosed | [Pile-FreeLaw]; [AMC 8 Problems and Solutions, AMC 10 Problems and Solution, and AIME Problems and Solutions] | [Nemotron 5.5] |\n| Vendor Terminal Bench-like Tasks from Mercor | Text | Undisclosed | [Terminal bench like tasks curated by the vendor] | [Undisclosed - purchased dataset] |\n| Turing Math Data Pack | Text | Undisclosed | [Turing Math Data Pack dataset] | [Undisclosed - purchased dataset] |\n| Synthetic Holdout, Skywork, DAPO, and Turing Math from GPT-5.5 | Text | Undisclosed | [DocQA-RL-1.6K]; [DAPO-Math-17k] | [GPT-5.5] |\n| Synthetic Long Context RL from QwenLong L1 and DocQA-RL-1.6K | Text | Undisclosed | [DocQA-RL-1.6K] | Undisclosed |\n| Synthetic Competitive Coding Gym Tasks | Text | Undisclosed | [NVCompetitiveCodingV1.1] | Undisclosed |\n| Synthetic Finance SEC Search Agent from GPT-OSS-120B and Qwen3 | Text | Undisclosed | [SEC filings from sec.gov] | [GPT-OSS-120B]; [Qwen3-235B-A22B-Instruct]; [Qwen3-4B-Instruct] |\n| Synthetic Structured Outputs from Qwen3-30B-A3B-Instruct-2507, Qwen3-30B-A3B-Thinking-2507, Qwen3-235B-A22B-Instruct-2507, and Qwen3-235B-A22B-Thinking-2507 | Text | Undisclosed | [Nemotron-RL-agent-structured-outputs-v1] | [Qwen3-30B-A3B-Instruct-2507]; [Qwen3-235B-A22B-Instruct-2507] |\n| Synthetic Long Context Equivalence Rule from Qwen3-235B-A22B-Thinking-2507 and DeepSeek-R1 | Text | Undisclosed | [Long-context SFT data] | [Qwen/Qwen3-235B-A22B-Thinking-2507]; [Deepseek-ai/DeepSeek-R1] |\n| Synthetic Science RL Data Blend from Qwen2.5-32B | Text | Undisclosed | [doubtnut]; [Pile-FreeLaw]; [Llama Nemotron Dataset]; [askfilo]; [EssentialAI/essential-web-v1.0]; [Vedantu]; [auxiliary_train]; [cdquestions.com]; [AMC 8 Problems and Solutions, AMC 10 Problems and Solution, and AIME Problems and Solutions]; [AAPT]; [ICHO-IPH0 Dataset]; [LIMO dataset (Less is More for Reasoning)] | [Qwen2.5-32B] |\n| Synthetic Abstention Data from Nemotron Super v3 | Text | Undisclosed | [Go abstention Dataset] | [nvidia/nvidia/nemotron-3-super-v3] |\n| Synthetic Chemistry Data from Nemotron Super v3 | Text | Undisclosed | [ChemData 700K] | [nvidia/nvidia/nemotron-3-super-v3] |\n| Synthetic Structured Outputs from Qwen3-30B-A3B-Instruct-2507, Qwen3-30B-A3B-Thinking-2507, Qwen3-235B-A22B-Instruct-2507, and Qwen3-235B-A22B-Thinking-2507 | Text | Undisclosed | [In-house data] | [GPT OSS 120B - Apache 2.0] |\n| Synthetic Tool Call Schema for RL | Text | Undisclosed | [In-house data] | [GPT OSS 120B - Apache 2.0] |\n| Synthetic Freeform Text Formatting from GPT-OSS-120B | Text | Undisclosed | [In-house data] | [GPT OSS 120B - Apache 2.0] |\n| Synthetic Citation Formatting from GPT-OSS-120B | Text | Undisclosed | [In-house data] | [GPT OSS 120B - Apache 2.0] |\n| Droid Harness Pivot Vendor Data | Text | Undisclosed | [Droid Harness Pivot vendor data] | Undisclosed |\n| Synthetic HotpotQA Training Data from Qwen3-235B | Text | Undisclosed | [HotpotQA] | [Qwen3-235B] |\n| Synthetic Natural Language Math Proofs from Nemotron 5.5 | Text | Undisclosed | [AMC8, AMC10, and AIME problem sets hosted on Art of Problem Solving]; [Pile-StackExchange] | [Nemotron 5.5] |\n| Synthetic Stack Overflow OpenQ | Text | Undisclosed | [Pile-FreeLaw] | Undisclosed |\n| Chemistry Ether0 Vendor Data | Text | Undisclosed | [Chemistry ether0 vendor data] | Undisclosed |\n| Synthetic Litmus-Bench Chemistry from ChEMBL | Text | Undisclosed | [ChEMBL]; [Nemo Gym RL dataset generated from ChEMBL with RDKit] | Undisclosed |\n| Synthetic ZINC Chemistry from Nemotron Super v3 | Text | Undisclosed | [ZINC] | [Nemotron Super v3] |\n| ARC-AGI Gym Environment | Text | Undisclosed | [ARC-AGI-2] | [ARC-AGI-2] |\n| Synthetic Agentic Search Tool-Use from DeepSeek-V3.2 | Text | Undisclosed | [Mercor Data] | [DeepSeek-V3.2] |\n| Synthetic Text-To-SQL | Text | Undisclosed | [In-house Text-to-SQL data] | [gpt-oss-120b] |\n| Dialog Memory Vendor Data | Text | Undisclosed | [Patronus external vendor agreement] | Undisclosed |\n| Synthetic Indirect Prompt Injection from Nemotron Super v3 and Qwen3-Next-80B-A3B-Instruct | Text | Undisclosed | [In-house indirect prompt injection data] | [nvidia/nemotron-3-super-v3, qwen/qwen3-next-80b-a3b-instruct.] |\n| Synthetic Malicious Code and Agentic Security | Text | Undisclosed | [In-house malicious-code / agentic-security data] | Undisclosed |\n| Synthetic Single-Step SWE Patch Selection | Text | Undisclosed | [SWE-Gym Dataset]; [SWE Bench Verified Benchmark] | [ground truth and task checks] |\n| Synthetic Natural Language Math Final Answers from Nemotron 5.5 | Text | Undisclosed | [AMC8, AMC10, and AIME problem sets hosted on Art of Problem Solving]; [Pile-StackExchange] | [nemotron 5.5] |\n| Synthetic Simple Math Prompts for Token Efficiency | Text | Undisclosed | [In-house simple math prompts] | Undisclosed |\n| Synthetic Abstention Data from Nemotron Super v3 | Text | Undisclosed | [CRAG] | [nvidia/nvidia/nemotron-3-super-v3] |\n| Synthetic Agentless SWE | Text | 242,536 | [SWE-Rebench-V2]; [SWEbench Training Set]; [R2E-Gym/R2E-Gym-Subset]; [SWE-Gym/SWE-Gym]; [SWE-Rebench] | [openai/gpt-oss-120b] |\n| Synthetic Agentic CUDA Traces from GLM-4.7 | Text | 2,276 | [Internal CUDA task data] | [GLM-4.7] |\n| Synthetic Math Proofs from DeepSeek-V3.2-Speciale | Text | 820,772 | [Nemotron-Math-Proofs-v1] | [SDG: DeepSeek-V3.2-Speciale]; [Filter: proof validation] |\n| Synthetic Multilingual SFT from DeepSeek-V3 | Text | 1,245,284 | [Nano v3 SFT data] | [DeepSeek-V3] |\n| Synthetic Agentic Code from gpt-oss-120b | Text | 109,086 | [NVAgenticCLIPrompts-v1]; [NVAgenticSkills-v1]; [NVAgenticCLIMultiTurnPrompts-v1] | [openai/gpt-oss-120b] |\n| Synthetic Agentic CLI and Web Skills from gpt-oss-120b | Text | 27,418 | [NVAgenticCLIPrompts-v1]; [NVAgenticSkills-v1]; [NVAgenticCLIMultiTurnPrompts-v1]; [NVAgenticCLIPrompts-Web-v1] | [openai/gpt-oss-120b] |\n| Synthetic Agentic Coding from gpt-oss-120b | Text | 160,531 | [NVAgenticCLIPrompts-v1]; [NVAgenticSkills-v1]; [NVAgenticCLIMultiTurnPrompts-v1]; [NVAgenticCLIPrompts-Web-v1] | [openai/gpt-oss-120b] |\n| Synthetic OpenCode Agentic Tasks from gpt-oss-120b | Text | 614,773 | [NVAgenticCLIPrompts-v1]; [NVAgenticSkills-v1]; [NVAgenticCLIMultiTurnPrompts-v1]; [NVAgenticCLIPrompts-Web-v1] | [openai/gpt-oss-120b] |\n| Synthetic ARC-AGI Ultra Data | Text | 192,016 | [ARC-AGI-2]; [arc dataset collection] | [ARC-AGI-2] |\n| Synthetic LiveCodeBench TIR from DeepSeek-R1-0528 | Text | 1,283,398 | [Nemotron-X training datasets] | [DeepSeek-R1-0528] |\n| Synthetic Verilog and SystemVerilog Code from DeepSeek-R1-0528 and GPT-OSS-120B | Text | 1,233,247 | [Verilog/SystemVerilog seed code] | [SDR: DeepSeek R1 0528 and GPT-OSS-120B]; [Filtering: Claude 4 Sonnet] |\n| Synthetic Aider Python Tasks from DeepSeek-R1-0528 | Text | 236,099 | [Exercism (GitHub Python)] | [Deepseek R1 0528] |\n| Synthetic Chat Reasoning-Off Data from GLM-5 | Text | 646,738 | [lmarena-ai/repochat-arena-preference-4k user prompts] | [Multi-turn conversations generated by GLM-5 with best-of-4 selection via Qwen3-Nemotron-235B-A22B-GenRM:] |\n| Synthetic Chat Reasoning-On Data from GLM-5 | Text | 644,286 | [lmarena-ai/repochat-arena-preference-4k user prompts]; [lmarena-ai/arena-expert-5k user prompts]; [lmarena-ai/arena-human-preference-55k user prompts]; [lmarena-ai/arena-human-preference-100k user prompts]; [lmarena-ai/arena-human-preference-140k user prompts] | [Multi-turn conversations generated by GLM-5 with best-of-4 selection via Qwen3-Nemotron-235B-A22B-GenRM:] |\n| Synthetic Multilingual Safety from Riva-Translate-4B-Instruct-v1.1 | Text | 132,067 | [Safety SFT Data: Ultra] | [nvidia/Riva-Translate-4B-Instruct-v1.1] |\n| Synthetic Science Reasoning Effort Medium | Text | 502,722 | [science-reasoning-effort-medium-v0] | Undisclosed |\n| Synthetic Telecom Tool-Use Trajectories from gpt-oss-120b | Text | 12,455 | [Existing Tau2 telecom trajectories originally generated with DeepSeek V3.2] | [gpt-oss-120b] |\n| Synthetic Terminal Bench Data from OpenReasoningv2 | Text | Undisclosed | [OpenCodeReasoningv2]; [OpenMathReasoning]; [nemo-swe-bench-repos]; [SWE-Rebench]; [SWE-Fixer-110K] | [OpenReasoningv2] |\n| Synthetic Tulu Instruction Following from DeepSeek-R1-0528 | Text | 105,361 | [Nemotron-X training datasets] | [DeepSeek-R1-0528] |\n| Synthetic SWE Unverified | Text | Undisclosed | [NVAgenticCLIPrompts-v1]; [NVAgenticSkills-v1]; [NVAgenticCLIMultiTurnPrompts-v1]; [NVAgenticCLIPrompts-Web-v1] | [gpt-oss-120b] |\n| Synthetic Instruction Following from gpt-oss-120b | Text | 151,988 | [IFEval]; [IFEvalG] | [gpt-oss-120b] |\n| Synthetic Identity Data from Qwen3-Next-80B-A3B-Instruct and Qwen3-235B-A22B-Instruct-2507 | Text | 25,992 | [Hand-written prompts] | [Qwen3-Next-80B-A3B-Instruct]; [Qwen3-235B-A22B-Instruct-2507] |\n| Synthetic Terminus Ultra Agentic Reasoning Blend | Text | 96,881 | [ARC-AGI-2]; [OpenCodeReasoningv2]; [OpenMathReasoning]; [SWE-Fixer-110K]; [SWE-Rebench]; [SWE-Smith] | [DeepSeek-V3.2]; [Qwen3-235B-A22B-Thinking-2507]; [Ring-1T]; [Kimi-K2.5]; [GLM-4.7-FP8]; [Qwen3-Next-80B-A3B-Thinking]; [gpt-oss-120b]; [Ministral-3-14B-Reasoning-2512]; [LM-4.5-Air-FP8] |\n| Synthetic STEM from Qwen3-235B-A22B-Thinking-2507 | Text | 1,174,694 | [IChO-IPhO-RL-v2]; [Physics-Big Dataset] | Undisclosed |\n| Translation Data from TAUS | Text | 1,618,055 | [TAUS proprietary dataset] | Undisclosed |\n| Synthetic Art of Problem Solving and Stack Exchange from gpt-oss-120b, Qwen2.5-32B-Instruct, and Goedel-Prover-V2-32B | Text | 860,469 | [Nemotron-Math-Proofs-v1] | [Goedel-Prover-V2-32B] |\n| Synthetic Art of Problem Solving and Stack Exchange from gpt-oss-120b, Qwen2.5-32B-Instruct, and Goedel-Prover-V2-32B | Text | 1,201,815 | [Upstream released math dataset]; [AoPS]; [StackOverflow / StackExchange] | [gpt-oss-120b] |\n| Synthetic Art of Problem Solving and Stack Exchange from gpt-oss-120b, Qwen2.5-32B-Instruct, and Goedel-Prover-V2-32B | Text | 1,296,676 | [Upstream released math dataset]; [AoPS]; [StackOverflow / StackExchange] | [gpt-oss-120b] |\n| Synthetic Instruction Following for RL | Text | Undisclosed | [WildChat-1M]; [LMSYS-340B-Eval Dataset]; [LMSYS-Chat-1M Prompts]; [IFEval]; [IFEvalG] | [Qwen/Qwen3-235B-A22B-Thinking-2507]; [gpt-oss-120b]; [Qwen3-235B-A22B-Instruct-2507] |\n| Synthetic Instruction Following for RL | Text | Undisclosed | [WildChat-1M]; [LMSYS-340B-Eval Dataset]; [LMSYS-Chat-1M Prompts]; [IFEval]; [IFEvalG] | [Qwen/Qwen3-235B-A22B-Thinking-2507]; [gpt-oss-120b]; [Qwen3-235B-A22B-Instruct-2507] |\n| Synthetic Multilingual Science and Code data from DeepSeek-R1, DeepSeek-R1-0528, Qwen2.5-32B-Instruct, and Qwen3-235B-A22B, translated with Qwen2.5-32B-Instruct and Qwen2.5-14B-Instruct | Text | Undisclosed | [Nano-V3 SFT Data (without tool call)] | [Qwen/Qwen2.5-14B-Instruct]; [Qwen/Qwen3-4B-Thinking-2507] |\n| Synthetic Search Graph Walk | Text | 6,977 | [Wikidata / Wikipedia KnowledgeBase] | [MiniMaxAI/MiniMax-M2] |\n| Synthetic Agentic Diverse Domains | Text | 281,537 | [Handwritten prompts (synthetic; no external seed data used)] | [SDG model: deepseek-ai/DeepSeek-V3.2, deepseek-ai/DeepSeek-R1-0528, Qwen/Qwen3-235B-A22B-Thinking-2507, Qwen/Qwen3-32B]; [Filtering model: openai/gpt-oss-120b, Qwen/Qwen3-32B, Qwen/Qwen3-235B-A22B-Instruct-2507] |\n| Synthetic Long Context from Qwen3-235B-A22B-Instruct-2507 | Text | 65,608 | [Long-context SFT seed blend (pre-training blend + nano-v1 post-training data)] | [Qwen/Qwen3-235B-A22B-Thinking-2507 and deepseek-ai/DeepSeek-R1] |\n| Synthetic Agentless SWE | Text | 209,976 | [SWE-Bench-Train]; [SWE-Fixer-Train]; [SWE-reBench]; [SWE-Smith] | [deepseek-ai/DeepSeek-R1-0528] |\n| Synthetic Nemotron Math SFT from DeepSeek-V3.2-Speciale | Text | 1,900,553 | [Nemotron-Math-v2 (AOPS and StackExchange-math problems)] | [DeepSeek-V3.2-Speciale] |\n| Synthetic Nemotron Math TIR from DeepSeek-V3.2 | Text | 1,789,258 | [Nemotron-Math-v2 (AOPS and StackExchange-math problems)] | [DeepSeek-V3.2] |\n| Synthetic SWE Unverified | Text | 27,911 | [NVAgenticCLIPrompts-v1]; [NVAgenticSkills-v1] | [gpt-oss-120b] |\n| Synthetic SWE Unverified | Text | 28,116 | [NVAgenticCLIPrompts-v1]; [NVAgenticSkills-v1] | [Qwen3-Coder-480B-A35B-Instruct] |\n| Synthetic NemoCascade OCR Distillation from gpt-oss-120b | Text | 682,864 | [Nemotron-X training datasets] | [gpt-oss-120b] |\n| Synthetic SWE Unverified | Text | 26,865 | [NVAgenticCLIPrompts-v1]; [NVAgenticSkills-v1] | [gpt-oss-120b]; [Qwen/Qwen3-Coder-480B-A35B-Instruct]; [GLM-4.7-Flash] |\n| Synthetic CUDA 100k | Text | 93,086 | [KernelBook]; [HuggingFace Transformers]; [FlashInfer] | [gpt-oss-120b]; [DeepSeek-R1-0528] |\n| Synthetic Science MCQ and QA Diversity from GPT-OSS and Kimi-K2 | Text | 30,358 | [doubtnut]; [Pile-FreeLaw]; [Llama Nemotron Dataset]; [askfilo]; [EssentialAI/essential-web-v1.0]; [Vedantu]; [auxiliary_train]; [cdquestions.com]; [AMC 8 Problems and Solutions, AMC 10 Problems and Solution, and AIME Problems and Solutions]; [AAPT]; [ICHO-IPH0 Dataset]; [LIMO dataset (Less is More for Reasoning)] | [GPT-OSS]; [Kimi-K2] |\n| Synthetic Science HLE with Python from GPT-OSS and Kimi-K2 | Text | 85,184 | [doubtnut]; [Pile-FreeLaw]; [Llama Nemotron Dataset]; [askfilo]; [EssentialAI/essential-web-v1.0]; [Vedantu]; [auxiliary_train]; [cdquestions.com]; [AMC 8 Problems and Solutions, AMC 10 Problems and Solution, and AIME Problems and Solutions]; [AAPT]; [ICHO-IPH0 Dataset]; [LIMO dataset (Less is More for Reasoning)] | [GPT-OSS]; [Kimi-K2] |\n| Synthetic Science Search and Python from GPT-OSS and Kimi-K2 | Text | 6,179 | [doubtnut]; [Pile-FreeLaw]; [Llama Nemotron Dataset]; [askfilo]; [EssentialAI/essential-web-v1.0]; [Vedantu]; [auxiliary_train]; [cdquestions.com]; [AMC 8 Problems and Solutions, AMC 10 Problems and Solution, and AIME Problems and Solutions]; [AAPT]; [ICHO-IPH0 Dataset]; [LIMO dataset (Less is More for Reasoning)] | [GPT-OSS]; [Kimi-K2] |\n| Synthetic Science Search from GPT-OSS and Kimi-K2 | Text | 32,554 | [doubtnut]; [Pile-FreeLaw]; [Llama Nemotron Dataset]; [askfilo]; [EssentialAI/essential-web-v1.0]; [Vedantu]; [auxiliary_train]; [cdquestions.com]; [AMC 8 Problems and Solutions, AMC 10 Problems and Solution, and AIME Problems and Solutions]; [AAPT]; [ICHO-IPH0 Dataset]; [LIMO dataset (Less is More for Reasoning)] | [GPT-OSS]; [Kimi-K2] |\n| Synthetic Finance Reasoning from GPT-OSS-120B and Qwen3-235B-A22B-Instruct-2507 | Text | 326,700 | [_SEC filings] | [GPT-OSS-120B, Qwen3-235B-A22B-Instruct-2507] |\n| Synthetic Science Diversity MCQ from GPT-OSS and Kimi-K2 | Text | 532,942 | [doubtnut]; [Pile-FreeLaw]; [Llama Nemotron Dataset]; [askfilo]; [EssentialAI/essential-web-v1.0]; [Vedantu]; [auxiliary_train]; [cdquestions.com]; [AMC 8 Problems and Solutions, AMC 10 Problems and Solution, and AIME Problems and Solutions]; [AAPT]; [ICHO-IPH0 Dataset]; [LIMO dataset (Less is More for Reasoning)] | [GPT-OSS]; [Kimi-K2] |\n| Synthetic Science Diversity OpenQ from GPT-OSS and Kimi-K2 | Text | 131,045 | [doubtnut]; [Pile-FreeLaw]; [Llama Nemotron Dataset]; [askfilo]; [EssentialAI/essential-web-v1.0]; [Vedantu]; [auxiliary_train]; [cdquestions.com]; [AMC 8 Problems and Solutions, AMC 10 Problems and Solution, and AIME Problems and Solutions]; [AAPT]; [ICHO-IPH0 Dataset]; [LIMO dataset (Less is More for Reasoning)] | [GPT-OSS]; [Kimi-K2] |\n| Synthetic Science Reasoning No-Tool from GPT-OSS and Kimi-K2 | Text | 2,085,600 | [doubtnut]; [Pile-FreeLaw]; [Llama Nemotron Dataset]; [askfilo]; [EssentialAI/essential-web-v1.0]; [Vedantu]; [auxiliary_train]; [cdquestions.com]; [AMC 8 Problems and Solutions, AMC 10 Problems and Solution, and AIME Problems and Solutions]; [AAPT]; [ICHO-IPH0 Dataset]; [LIMO dataset (Less is More for Reasoning)] | [GPT-OSS]; [Kimi-K2] |\n| Synthetic Long Context from Qwen3-235B-A22B-Instruct-2507 | Text | 62,333 | [Long-context SFT data: lc_nothink 256k] | [Qwen/Qwen3-235B-A22B-Thinking-2507 and deepseek-ai/DeepSeek-R1] |\n| Synthetic Long Context from Qwen3-235B-A22B-Instruct-2507 | Text | 49,698 | [Long-context SFT data: MRCR 200k] | [Qwen/Qwen3-235B-A22B-Thinking-2507 and deepseek-ai/DeepSeek-R1] |\n| Synthetic Text-To-SQL | Text | 96,564 | [Undisclosed - no seed data listed] | [gpt-oss-120b] |\n| Synthetic Long Context from Qwen3-235B-A22B-Instruct-2507 | Text | 397,538 | [Long-context SFT data: RULER 256k] | [Qwen/Qwen3-235B-A22B-Thinking-2507 and deepseek-ai/DeepSeek-R1] |\n| Synthetic SWE Unverified | Text | 27,960 | [NVAgenticCLIPrompts-v1]; [NVAgenticSkills-v1] | [gpt-oss-120b]; [Qwen/Qwen3-Coder-480B-A35B-Instruct]; [GLM-4.7-Flash] |\n| Synthetic SWE Unverified | Text | 24,632 | [NVAgenticCLIPrompts-v1]; [NVAgenticSkills-v1] | [gpt-oss-120b]; [Qwen/Qwen3-Coder-480B-A35B-Instruct]; [GLM-4.7-Flash] |\n| Synthetic Long Context from Qwen3-235B-A22B-Instruct-2507 | Text | 49,902 | [Long-context SFT data] | [Qwen/Qwen3-235B-A22B-Thinking-2507 and deepseek-ai/DeepSeek-R1] |\n| Synthetic Tool Call Schema for RL | Text | 469,983 | [UltraTool]; [ToolEyes]; [AutoTools]; [API-Bank]; [Nemotron-Personas-USA]; [Salesforce xLAM function-calling]; [Glaive function-calling-v2]; [Agent-Ark/Toucan-1.5M] | [DeepSeek-V3.2]; [GLM-4.6]; [gpt-oss-120b]; [Kimi-K2-Instruct] |\n| Synthetic Tool Call Schema for RL | Text | 707,967 | [UltraTool]; [ToolEyes]; [AutoTools]; [API-Bank]; [Nemotron-Personas-USA]; [Salesforce xLAM function-calling]; [Glaive function-calling-v2]; [Agent-Ark/Toucan-1.5M] | [DeepSeek-V3.2]; [GLM-4.6]; [gpt-oss-120b]; [Kimi-K2-Instruct] |\n| Synthetic Long Context from Qwen3-235B-A22B-Instruct-2507 | Text | 52,630 | [AALCR seed blend: SEC Filings]; [CC]; [Wikipedia]; [FinePDFs]; [ArXiv]; [Pile-NIH ExPorter]; [BioRxiv]; [PMC Article]; [USPTO Backgrounds]; [peS20]; [Global Regulations]; [CORE]; [Gutenberg (PG-19)]; [DOAB CC-BY]; [NDLTD]; [Amps]; [StackExchange]; [MathPile]; [Numinas] | [Qwen3-30B-A3B] |\n| Synthetic Safety from gemma-3-4b-it, Nemotron-Nano-9B-v2, and gpt-oss-120b | Text | 44,091 | [Safety SFT Data] | [google/gemma-3-4b-it]; [Nemotron-Nano-9B-v2]; [gpt-oss-120b] |\n\n## Language Distribution in Post-Training\n\nFor our post-training recipe, we focused on the following languages in addition to English: French, Spanish, Italian, German, Japanese, Korean, Hindi, Korean, Brazilian Portuguese, and Chinese.\n\nThose languages were represented in the form of multilingual reasoning and translation tasks.\n\nThe following table depicts our sample distribution.\n\n| Language | Size |\n|---|---:|\n| English | 8.6M |\n| Italian | 138k |\n| German | 138k |\n| Spanish | 138k |\n| French | 138k |\n| Japanese | 138k |5\n| Chinese | 138k |\n| Hindi | 138k |\n| Korean | 138k |\n| Brazilian Portuguese | 138k |\n\n## Evaluation Datasets:\n\n**Data Collection Method by dataset**
\n* Hybrid: Automated, Human, Synthetic\n\n**Labeling Method by dataset**
\n* Hybrid: Automated, Human, Synthetic\n\n**Properties:** This corpus comprises a mix of high-quality standard benchmarks and test suites for modern agentic AI as outlined in the benchmark section of the model card.\n\n## Testing Datasets:\n\n**Data Collection Method by dataset**
\n* Hybrid: Automated, Human, Synthetic\n\n**Labeling Method by dataset**
\n* Hybrid: Automated, Human, Synthetic\n\n**Properties:** This corpus comprises a mix of high-quality standard benchmarks and test suites for modern agentic AI as outlined in the benchmark section of the model card.\n\n

\n\n## Inference\n\n* **Acceleration Engine:** PyTorch\n* **Test Hardware:**\n * NVIDIA Hopper\n * H100\n * H200\n * NVIDIA Grace Blackwell\n * GB200\n * GB300\n * NVIDIA Blackwell\n * B200\n * B300\n\n## Ethical Considerations\n\nNVIDIA believes Trustworthy AI is a shared responsibility and we have established policies and practices to enable development for a wide array of AI applications. When downloaded or used in accordance with our terms of service, developers should work with their internal model team to ensure this model meets requirements for the relevant industry and use case and addresses unforeseen product misuse.\n\nWe advise against circumvention of any provided safety guardrails contained in the Model without a substantially similar guardrail appropriate for your use case. For more details: [Safety](./safety.md) and [Explainability](./explainability.md) Subcards.\n\nFor more detailed information on ethical considerations for this model, please see the Model Card++ [Bias](./bias.md), and [Privacy](./privacy.md) Subcards.\n\nPlease report model quality, risk, security vulnerabilities or NVIDIA AI Concerns [here](https://www.nvidia.com/en-us/support/submit-security-vulnerability/).\n\n## **Get Help**\n\n### NVIDIA Developer Community Forum\nGet access to community knowledge base articles and support cases (https://forums.developer.nvidia.com/)\n\n### Enterprise Support\nGet access to knowledge base articles and support cases or [submit a ticket](https://www.nvidia.com/en-us/data-center/products/ai-enterprise-suite/support/).\n\n\n## Citation\n\n```bibtex\n@misc{nvidia_nemotron_3_ultra_2026,\n title = {Nemotron 3 Ultra: Open, Efficient Mixture-of-Experts Hybrid Mamba-Transformer Model for Agentic Reasoning},\n author = {{NVIDIA}},\n year = {2026},\n url = {https://research.nvidia.com/labs/nemotron/files/NVIDIA-Nemotron-3-Ultra-Technical-Report.pdf},\n note = {White Paper}\n}\n```","displayName","NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16",["D",1785274063103],"publisher","NVIDIA","latestVersion","versions","collectionCounts","imageUrl","/api/image/nim/nvidia/models/nemotron-3-ultra-550b-a55b","rawImageUrl","","canGuestDownload","latestVersionSizeInBytes",107,"licenseTerms",{"_300":301,"_302":303,"_304":305,"_306":30,"_307":308},"isMultinodeEnabled","isMultiArchitecture","bias","| Field | Response |\n| :---- | :---- |\n| Participation considerations from adversely impacted groups [protected classes](https://www.senate.ca.gov/content/protected-classes) in model design and testing: | None |\n| Bias Metric (If Measured): | [BBQ Accuracy Scores in Ambiguous Contexts](https://github.com/nyu-mll/BBQ/) |\n| Which characteristic (feature) show(s) the greatest difference in performance?: | The model shows high variance in the characteristics when it is used with a high temperature. |\n| Measures taken to mitigate against unwanted bias: | Supervised Fine-Tuning (SFT) and Reinforcement Learning from Human Feedback (RLHF) employed to calibrate the model’s reasoning capabilities to maintain logical consistency and appropriate complexity when interacting with or interpreting data from diverse age demographics. |\n| If using internal data, description of methods implemented in data acquisition or processing, if any, to address the prevalence of identifiable biases in the training, testing, and validation data: | The training datasets contain a large amount of synthetic data generated by LLMs. We manually curated prompts. |\n| Tools used to assess statistical imbalances and highlight patterns that may introduce bias into AI models: | [BBQ](https://github.com/nyu-mll/BBQ/) |\n| Tools used to assess statistical imbalances and highlight patterns that may introduce bias into AI models: | These datasets, such as web-scraped finance reasoning data derived from SEC EDGAR filings, science and math problem datasets, OpenResearcher/source-document datasets, Common Crawl, CC-News, Wikimedia, and long-context document datasets, do not collectively or exhaustively represent all demographic groups (and proportionally therein). For instance, these datasets do not contain explicit mentions of demographic classes such as age, gender, or ethnicity in approximately 97% to 99.9% of finance reasoning samples and in over 85% of samples across the broader assessed datasets. In the subset where such terms are present, these datasets contain notable representational skews. For example, ethnicity mentions are often dominated by Middle Eastern contexts (found in finance documents) or \"White,\" \"Two or more,\" and \"Black or African American\" as the most frequent ethnic identifiers, while references categorized as male-only significantly outnumber those categorized as female-only. Furthermore, gender is explicitly mentioned in approximately 12% of samples across the broader dataset assessment, yet in only 0.9% of finance-specific samples. Dataset-level results vary by source type, with long-context/source-document datasets containing higher explicit demographic mention rates compared to certain web-scraped sources. To mitigate these imbalances, we recommend considering evaluation techniques such as bias audits, fine-tuning with demographically balanced datasets, and mitigation strategies such as counterfactual data augmentation to align with the desired model behavior. This evaluation used a 3,000-sample subset per dataset, identified as the optimal threshold for maximizing embedder accuracy. |\n| Unwanted Bias Testing: | Constrained to English-language inputs. Multi-lingual parity is not currently claimed or guaranteed. |","explainability","| Field | Response |\n| :---- | :---- |\n| Intended Task/Domain: | Text generation, reasoning, and chat |\n| Model Type: | Text-to-text Mamba2-Transformer Hybrid |\n| Intended Users: | Generative AI creators working with conversational AI models and image content. |\n| Output: | Text |\n| Tools used to evaluate datasets to identify synthetic data and ensure data authenticity. | We used a Gemma-3 4B-based filtering model fine-tuned on [Nemotron Content Safety Dataset v2](https://huggingface.co/datasets/nvidia/Aegis-AI-Content-Safety-Dataset-2.0) to ensure the quality of synthetic data. |\n| Describe how the model works: | Generates text by predicting the next word or token based on the context provided in the input sequence using multiple self-attention layers. |\n| Name the adversely impacted groups this has been tested to deliver comparable outcomes regardless of: | Age, Disability Status, Gender Identity, Nationality, Physical Appearance, Ethnicity, Socioeconomic Status, Sexual Orientation, Religion |\n| Technical Limitations & Mitigation: | This model performs particularly well in instruction following regimes, as such may be strongly influenced by untrusted inputs and should be paired with appropriate guardrails and data filtering to better align use-case behaviors when exposed to such data. |\n| Verified to have met prescribed NVIDIA quality standards: | Yes |\n| Performance Metrics: | Accuracy, Throughput, and User-side throughput |\n| Potential Known Risks: | The model was optimized explicitly for instruction following and as such is more susceptible to prompt injection and jailbreaking in various forms as a result of its instruction tuning. This means that the model should be paired with additional rails or system filtering to limit exposure to instructions from malicious sources -- either directly or indirectly by retrieval (e.g. via visiting a website) -- as they may yield outputs that can lead to harmful, system-level outcomes up to and including remote code execution in agentic systems when effective security controls including guardrails are not in place. The model may generate answers that may be inaccurate, omit key information, include irrelevant or redundant text, or produce socially unacceptable or undesirable text, even if the prompt itself does not include anything explicitly offensive. |\n| Licensing: | Use of this model is governed by the [OpenMDW License Agreement, version 1.1](https://raw.githubusercontent.com/OpenMDW/OpenMDW/refs/heads/main/1.1/LICENSE.OpenMDW-1.1) (OpenMDW-1.1). |","privacy","| Privacy Information |\n| :--- |\n| Nemotron 3 Ultra was trained on large-scale publicly available data that may contain images, audio-video, and text relating to people. NVIDIA collected and used this data in compliance with applicable data protection and privacy laws. This model was not designed to derive insights or otherwise learn from any personal data contained in the datasets. |\n| NVIDIA uses a combination of filters, data minimization techniques, and other guardrails to help prevent personal data from being recited by our models. We employ automated tools and data processing techniques during pre-training or training to identify and filter certain categories of personal data. |\n| Please review NVIDIA's [Privacy Policy](https://www.nvidia.com/en-us/about-nvidia/privacy-policy/) for more information. |","safetyAndSecurity","| Field | Response |\n| :---- | :---- |\n| Model Application Field(s): | Chat, Instruction Following, Chatbot Development, Code Generation, Reasoning, Customer Service |\n| Describe the life critical impact (if present). | Not Applicable |\n| Description of methods implemented in data acquisition or processing, if any, to address other types of potentially harmful data in the training, testing, and validation data: | We used a guard model for content safety to exclude potentially harmful data from training. |\n| Description of any methods implemented in data acquisition or processing, if any, to address illegal or harmful content in the training data, including, but not limited to, child sexual abuse material (CSAM) and non-consensual intimate imagery (NCII) | We used a Gemma-3 4B-based guard model trained on [Nemotron Content Safety Dataset v2](https://huggingface.co/datasets/nvidia/Aegis-AI-Content-Safety-Dataset-2.0) for content safety to exclude potentially illegal or harmful content from the training. |\n| Use Case Restrictions: | Use of this model is governed by the [OpenMDW License Agreement, version 1.1](https://raw.githubusercontent.com/OpenMDW/OpenMDW/refs/heads/main/1.1/LICENSE.OpenMDW-1.1) (OpenMDW-1.1).|\n| Model and dataset restrictions: | The Principle of least privilege (PoLP) is applied limiting access for dataset generation and model development. Restrictions enforce dataset access during training, and dataset license constraints adhered to. |\n| This AI model was developed based on our policies to ensure responsible data handling and risk mitigation. The datasets used for training have been scanned for harmful content and illegal content, consistent with our policies including scanning for Child Sexual Abuse Material (CSAM). Ongoing review and monitoring mechanisms are in place based on our policies and to maintain data integrity. | True. We use [Nemotron Content Safety Dataset V2](https://huggingface.co/datasets/nvidia/Aegis-AI-Content-Safety-Dataset-2.0) and an internal safety dataset specialized for minority sexuality for content safety evaluation to ensure the safety of this model. |","featureLabels",[291,292],"displayLabels",[277,278,279],"policyLabels",[],"systemLabels",[],"products",[],"hasNimMetadata",{"_239":289,"_281":289,"_282":290,"_284":285},{"_239":287,"_281":287,"_282":288,"_284":285},{"_239":280,"_281":280,"_282":283,"_284":285},"NSPECT-BJ7P-NJXS","value","filter",{"_286":285,"_281":280},"category","label","field","ingested",{"_286":285,"_281":287},"False",{"_286":285,"_281":289},{"_239":298,"_281":298,"_282":299,"_284":296},{"_239":293,"_281":294,"_282":295,"_284":296},"NVIDIA NIM","nimmcro_nvidia_nim",{"_286":297,"_281":294},"feature","nvidia_nim","NVIDIA AI Enterprise Supported",{"_286":285,"_281":298},"governingTerms","

GOVERNING DOWNLOAD TERMS: Use of the model is governed by the OpenMDW License Agreement, version 1.1.

","licenseId","NVIDIA proprietary","licenseVersion","v1.0","needsAcceptance","configuredLicenseTerms",{"_309":310,"_311":312},"hosting",{"_313":314,"_315":327,"_317":328},"download",{"_313":314,"_315":316,"_317":318},"artifactType","model","tags",[319],"generatedHtml","

GOVERNING DOWNLOAD TERMS: Use of the model is governed by the OpenMDW License Agreement, version 1.1.

",{"_230":320,"_284":321,"_322":323,"_324":325,"_285":326},"openmdw-license-agreement-version-1-1","terms","tag","OpenMDW-1.1","url","https://github.com/OpenMDW/OpenMDW/blob/main/1.1/LICENSE.OpenMDW-1.1","OpenMDW License Agreement, version 1.1",[329],"

GOVERNING HOSTING TERMS: Use of the model is governed by the OpenMDW License Agreement, version 1.1.

",{"_230":320,"_284":321,"_322":323,"_324":325,"_285":326},{"_105":1309,"_107":1310},"ngcApps",[1262],"csaApiBaseUrl","https://www.nvidia.com/ai-agents/csa-agent/dcsm","datadogRum",{"_1243":1244,"_1245":1246,"_1247":1248,"_1249":1250,"_1251":341,"_137":367,"_1252":1253,"_1254":55,"_1255":84,"_1256":84,"_1257":1258,"_1259":84,"_1260":1261},"amplitude",{"_1240":1241,"_880":-7,"_1242":30},"disableJs","environment","prod","featureFlagsBlob","JTdCJTIydGVzdEZsYWclMjIlM0FmYWxzZSUyQyUyMnNob3dTZWFyY2hMYWJlbFNldEZpbHRlcnMlMjIlM0FmYWxzZSUyQyUyMmVuYWJsZUNzYVdpZGdldCUyMiUzQWZhbHNlJTJDJTIyZW5hYmxlUHJvZHVjdEdyb3VwaW5nQnlTdWJzY3JpcHRpb24lMjIlM0FmYWxzZSUyQyUyMmNvbnRhaW5lclNjYW5uaW5nUmVkZXNpZ25NMiUyMiUzQWZhbHNlJTdE","featureFlagsGuarded","isDevLikeEnv","isPlaywright","isMobileDevice","categorizedOrgs",{"_1234":1235,"_1236":1237,"_1238":1239},"currentOrgContext","currentContextRoles",[],"enterpriseOrgs",[],"manageableExternalUserGroups",[],"ngcApiOrigin","ngcHelmOrigin","origin","https://catalog.ngc.nvidia.com","usesMocks","theme","year",2026,"now",1789019917452,"1.15.0","defaultSearchSuggestions",{"_1120":250,"_1121":1122,"_1123":1124,"_1125":129},"announcements",[],"consentState",{"_406":1110,"_1111":84,"_1112":30,"_1113":30,"_1114":1115},"consentConfig",{"_388":389,"_390":391,"_392":393,"_394":395,"_396":397,"_398":399,"_400":401,"_402":403,"_404":405,"_406":407},"showContactAdmin",1789019917454,[],[],[381,382,383,384,385,386],{"_44":58,"_46":68,"_48":100,"_50":-5,"_52":-5,"_54":55},{"_44":58,"_46":66,"_48":100,"_50":-5,"_52":-5,"_54":55},{"_44":58,"_46":64,"_48":100,"_50":-5,"_52":-5,"_54":55},{"_44":58,"_46":59,"_48":100,"_50":-5,"_52":-5,"_54":55},{"_44":56,"_46":57,"_48":49,"_50":51,"_52":387,"_54":55},{"_44":45,"_46":47,"_48":100,"_50":-5,"_52":-5,"_54":55},5,"mainInfoText","We and our third-party partners (including social media, advertising, and analytics partners) use cookies and other tracking technologies to collect, store, monitor, and process certain information about you when you visit our website. The information collected might relate to you, your preferences, or your device. We use that information to make the site work, analyze performance and traffic on our website, provide a more personalized web experience, and assist in our marketing efforts.

Under certain privacy laws, you have the right to direct us not to \"sell\" or \"share\" your personal information for targeted advertising. To opt-out of the \"sale\" and \"sharing\" of personal information through cookies, you must opt-out of optional cookies using the toggles below. To opt out of the \"sale\" and \"sharing\" of data collected by other means (e.g., online forms) you must also update your data sharing preferences through the NVIDIA Preference Center.

Click on the different category headings below to find out more and change the settings according to your preference. You cannot opt out of Required Cookies as they are deployed to ensure the proper functioning of our website (such as prompting the cookie banner and remembering your settings, etc.). By clicking \"Save and Accept\" or \"Decline All\" at the bottom, you consent to the use of cookies and other tools as described in our Cookie Policy in accordance with your settings and accept our Terms of Service (which contains important waivers). For more information about our privacy practices, please see our Privacy Policy.","mainText","Cookie Settings","alertNoticeText",{"_1102":1103,"_1104":1105,"_1106":1107,"_1108":1109},"bannerAcceptLabel","Accept All","bannerRejectLabel","Reject Optional","bannerSettingsLabel","Manage Settings","preferenceSaveLabel","Save Choices","preferenceRejectLabel","Decline All","alwaysActiveText","Always Active","groups",[408,409,410,411,412],{"_413":792,"_76":793,"_166":794,"_417":84,"_418":16,"_420":795,"_422":796},{"_413":695,"_76":696,"_166":697,"_417":30,"_418":53,"_420":698,"_422":699},{"_413":602,"_76":603,"_166":604,"_417":30,"_418":605,"_420":606,"_422":607},{"_413":424,"_76":425,"_166":426,"_417":30,"_418":387,"_420":427,"_422":428},{"_413":414,"_76":415,"_166":416,"_417":30,"_418":419,"_420":421,"_422":423},"optanonGroupId","C0005","Social Media Cookies","These cookies are set by a range of social media services that we have added to the site to enable you to share our content with your friends and networks. They are capable of tracking your browser across other sites and building up a profile of your interests. This may impact the content and messages you see on other websites you visit. If you do not allow these cookies you may not be able to use or see these sharing tools.","alwaysActive","order",11,"cookies",[],"hosts",[],"C0004","Advertising Cookies","These cookies record your visit to our websites, the pages you have visited and the links you have followed to influence the advertisements that you see on other websites. These cookies and the information they collect may be managed by other companies, including our advertising partners, and may be used to build a profile of your interests and show you relevant advertising on other sites. We and our advertising partners will use this information to make our websites and the advertising displayed on it, more relevant to your interests.",[544,545,546,547,548,549,550,551,552,553,554,555,556,557,558,559,560,561,562,563,564,565],[429,430,431,432,433,434,435,436,437,438,439,440,441,442,443,444,445],{"_446":539,"_239":539,"_166":250,"_420":540},{"_446":535,"_239":535,"_166":250,"_420":536},{"_446":530,"_239":530,"_166":250,"_420":531},{"_446":526,"_239":526,"_166":250,"_420":527},{"_446":521,"_239":521,"_166":250,"_420":522},{"_446":516,"_239":516,"_166":250,"_420":517},{"_446":508,"_239":508,"_166":250,"_420":509},{"_446":500,"_239":500,"_166":250,"_420":501},{"_446":495,"_239":495,"_166":250,"_420":496},{"_446":490,"_239":490,"_166":250,"_420":491},{"_446":485,"_239":485,"_166":250,"_420":486},{"_446":480,"_239":480,"_166":250,"_420":481},{"_446":475,"_239":475,"_166":250,"_420":476},{"_446":468,"_239":468,"_166":250,"_420":469},{"_446":463,"_239":463,"_166":250,"_420":464},{"_446":456,"_239":456,"_166":250,"_420":457},{"_446":447,"_239":447,"_166":250,"_420":448},"hostName","amazon-adsystem.com",[449,450],{"_76":455,"_452":447,"_453":30,"_166":454},{"_76":451,"_452":447,"_453":30,"_166":454},"ad-privacy","host","isSession","This domain is owned by online retailer Amazon and is used as part of its affiliate marketing programme.","ad-id","doubleclick.net",[458,459],{"_76":462,"_452":456,"_453":30,"_166":461},{"_76":460,"_452":456,"_453":30,"_166":461},"IDE","This domain is owned by Doubleclick (Google). The main business activity is: Doubleclick is Googles real time bidding advertising exchange","test_cookie","bing.com",[465],{"_76":466,"_452":463,"_453":30,"_166":467},"MUID","This domain is owned by Mircosoft - it is the site for the search engine Bing.","assets.adobedtm.com",[470,471],{"_76":474,"_452":468,"_453":30,"_166":473},{"_76":472,"_452":468,"_453":30,"_166":473},"kndctr_F207D74D549850760A4C98C6_AdobeOrg_*","If you know anything about this host and its uses please add your contributions here. ","kndctr_F207D74D549850760A4C98C6_AdobeOrg_consent","tiktok.com",[477],{"_76":478,"_452":475,"_453":30,"_166":479},"_ttp","TikTok is a social media platform owned by ByteDance, known for short-form video sharing. The service is used to create and share videos, and it utilizes cookies for various purposes including advertising and analytics.","t.co",[482],{"_76":483,"_452":480,"_453":30,"_166":484},"muc_ads","This domain is owned by Twitter. It is used as a URL shortening service, which allows Twitter to collect data on how links are shared.","snowplow.com",[487],{"_76":488,"_452":485,"_453":30,"_166":489},"sp","Testing","google.com",[492],{"_76":493,"_452":490,"_453":30,"_166":494},"NID","This domain is owned by Google Inc. Although Google is primarily known as a search engine, the company provides a diverse range of products and services. Its main source of revenue however is advertising. Google tracks users extensively both through its own products and sites, and the numerous technologies embedded into many millions of websites around the world. It uses the data gathered from most of these services to profile the interests of web users and sell advertising space to organisations based on such interest profiles as well as aligning adverts to the content on the pages where its customer's adverts appear.","linkedin.com",[497],{"_76":498,"_452":495,"_453":30,"_166":499},"bcookie","This domain is owned by LinkedIn, the business networking platform. It typically acts as a third party host where website owners have placed one of its content sharing buttons in their pages, although its content and services can be embedded in other ways. Although such buttons add functionality to the website they are on, cookies are set regardless of whether or not the visitor has an active Linkedin profile, or agreed to their terms and conditions. For this reason it is classified as a primarily tracking/targeting domain.","twitter.com",[502,503],{"_76":506,"_452":500,"_453":30,"_166":507},{"_76":504,"_452":500,"_453":30,"_166":505},"personalization_id","This domain is owned by Twitter. The main business activity is: Social Networking Services. Where twitter acts as a third party host, it collects data through a range of plug-ins and integrations, that is primarily used for tracking and targeting.","guest_*","This domain is owned by X (formerly Twitter). The main business activity is: Social Networking Services. Where X (formerly Twitter) acts as a third party host, it collects data through a range of plug-ins and integrations, that is primarily used for tracking and targeting.","youtube.com",[510,511],{"_76":514,"_452":508,"_453":30,"_166":515},{"_76":512,"_452":508,"_453":30,"_166":513},"VISITOR_INFO1_LIVE","This cookie is used as a unique identifier to track viewing of videos","__Secure-*","YouTube is a Google owned platform for hosting and sharing videos. YouTube collects user data through videos embedded in websites, which is aggregated with profile data from other Google services in order to display targeted advertising to web visitors across a broad range of their own and other websites.","demdex.net",[518],{"_76":519,"_452":516,"_453":30,"_166":520},"demdex","This cookie helps Adobe Audience Manger perform basic functions such as visitor identification, ID synchronization, segmentation, modeling, reporting, etc.","im-apps.net",[523],{"_76":524,"_452":521,"_453":30,"_166":525},"receive-cookie-deprecation","This domain is owned by im-apps.net which provides information and resources about various instant messaging (IM) applications to help users make informed decisions based on their communication needs.","influ2.com",[528],{"_76":529,"_452":526,"_453":30,"_166":489},"R","6sc.co",[532],{"_76":533,"_452":530,"_453":30,"_166":534},"6suuid","This domain is associated with the company 6Sense Insights, Inc., which provides predictive analytics and account engagement services. It is primarily used for tracking and targeting in marketing automation and advertising.","hushly.com",[537],{"_76":538,"_452":535,"_453":30,"_166":489},"_hly_gvid","t.contentsquare.net",[541],{"_76":542,"_452":539,"_453":84,"_166":543},"_cs_*","This is a session cookie. This is a pattern type cookie where a random 13-digit number is appended to the prefix _cs_.",{"_76":600,"_452":567,"_453":30,"_166":601},{"_76":598,"_452":567,"_453":30,"_166":599},{"_76":596,"_452":567,"_453":30,"_166":597},{"_76":478,"_452":567,"_453":30,"_166":595},{"_76":572,"_452":567,"_453":84,"_166":574},{"_76":594,"_452":567,"_453":30,"_166":250},{"_76":593,"_452":570,"_453":30,"_166":250},{"_76":591,"_452":567,"_453":30,"_166":592},{"_76":589,"_452":567,"_453":30,"_166":590},{"_76":588,"_452":567,"_453":30,"_166":250},{"_76":586,"_452":567,"_453":30,"_166":587},{"_76":572,"_452":585,"_453":84,"_166":574},{"_76":583,"_452":570,"_453":30,"_166":584},{"_76":583,"_452":573,"_453":30,"_166":584},{"_76":542,"_452":567,"_453":30,"_166":543},{"_76":581,"_452":567,"_453":30,"_166":582},{"_76":488,"_452":579,"_453":30,"_166":580},{"_76":577,"_452":567,"_453":30,"_166":578},{"_76":575,"_452":567,"_453":30,"_166":576},{"_76":572,"_452":573,"_453":84,"_166":574},{"_76":569,"_452":570,"_453":30,"_166":571},{"_76":566,"_452":567,"_453":84,"_166":568},"AMCVS_xxxxxAdobeOrg","nvidia.com","This is a pattern type cookie name associated with Adobe Marketing Cloud. It stores a unique visitor identifier, and uses an organisation identifier to allow a company to track users across their domains and services.This is a pattern type cookie name associated with Adobe Marketing Cloud. It stores a unique visitor identifier, and uses an organisation identifier.","_rdt_pn","www.nvidia.com","This cookie name is associated with www.mind.org.uk which is used for marketing purposes.","AMP_*","build.nvidia.com","Supports Amplitude product analytics.","_gcl_au","Used by Google AdSense for experimenting with advertisement efficiency across websites using their services","_mkto_trk","This cookie is associated with an email marketing service provided by Marketo. This tracking cookie allows a website to link visitor behaviour to the recipient of an email marketing campaign, to measure campaign effectiveness.","investor.nvidia.com","This cookie is associated with Eventbrite and is used to deliver content tailored to the end user's interests and improve content creation. This cookie is also used for event-booking purposes.","_gat_gtag_UA_140825383_2","Google Analytics","_rdt_uuid","This cookie is set by Reddit and is used for tracking the views of embedded videos and performance of the advertisements.","catalog.ngc.nvidia.com","_uetsid","This cookie is used by Bing to determine what ads should be shown that may be relevant to the end user perusing the site. ","kndctr_F207D74D549850760A4C98C6_AdobeOrg_bc_session_id","ttcsid_XXXXXXXXXXXXXXXXXXXXCategory","This cookie name is associated with TikTok which is used to track your activity on the platform, including how you interact with their embedded services on other websites.This cookie is associated with TikTok. It helps track your activity on the platform, including how you interact with their embedded services on other websites.","_fbp","Used by Facebook to deliver a series of advertisement products such as real time bidding from third party advertisers","_zitok","_twpid","This cookie is typically associated with the use of Taptap Digital, a company specializing in digital advertising. It is used for tracking and ad personalization purposes.","_tt_enable_cookie","This cookie name is associated with TikTok which is used to collect data about behavior and purchases on the website and to measure the effect of advertising. It is also used to track website activity to help optimize advertising efforts.","AMCV_xxxxxAdobeOrg","This cookie is used to identify a unique visitor for tracking purposes by Adobe Experience Cloud.This is a pattern type cookie name associated with Adobe Marketing Cloud. It stores a unique visitor identifier, and uses an organisation identifier to allow a company to track users across their domains and services.","ttcsid","This cookie name is associated with TikTok which is used to enable TikTok Ads session tracking by assigning a unique session ID to users who click on TikTok ads, allowing for the attribution of user behavior and conversions within the same session.This cookie is associated with TikTok. It helps track your activity on the platform, including how you interact with their embedded services on other websites.","C0003","Personalization Cookies","These cookies collect data about how you have interacted with our website to help us improve your web experience, such as which pages you have visited. These cookies may store a unique ID so that our system will remember you when you return. They may be set by us or by third party providers whose services we have added to our pages. These cookies enable us to provide enhanced website functionality and personalization as well as make the marketing messages we send to you more relevant to your interests. If you do not allow these cookies, then some or all of these services may not function properly.",4,[637,638,639,640,641,642,643,644,645,646,647,648,649,650,651,652,653,654,655,656,657,658,659,660,661],[608,609,610,611,612],{"_446":521,"_239":521,"_166":250,"_420":632},{"_446":495,"_239":495,"_166":250,"_420":627},{"_446":622,"_239":622,"_166":250,"_420":623},{"_446":617,"_239":617,"_166":250,"_420":618},{"_446":613,"_239":613,"_166":250,"_420":614},"nvidia.partners",[615],{"_76":616,"_452":613,"_453":30,"_166":489},"edge","app.hushly.com",[619],{"_76":620,"_452":617,"_453":30,"_166":621},"*CORS","This domain is associated with Hushly, a platform that provides lead conversion solutions for content marketing by using personalization and content engagement strategies.","github.com",[624],{"_76":625,"_452":622,"_453":30,"_166":626},"_octo","This domain is owned by Github. The main business activity is: Source code repository and collaboration platform",[628,629],{"_76":631,"_452":495,"_453":30,"_166":499},{"_76":630,"_452":495,"_453":30,"_166":499},"lidc","li_gc",[633,634],{"_76":636,"_452":521,"_453":30,"_166":525},{"_76":635,"_452":521,"_453":30,"_166":525},"imid_*","imid",{"_76":662,"_452":570,"_453":30,"_166":664},{"_76":693,"_452":567,"_453":30,"_166":694},{"_76":662,"_452":692,"_453":30,"_166":664},{"_76":690,"_452":567,"_453":84,"_166":691},{"_76":688,"_452":567,"_453":30,"_166":689},{"_76":687,"_452":670,"_453":84,"_166":250},{"_76":662,"_452":666,"_453":30,"_166":664},{"_76":685,"_452":567,"_453":30,"_166":686},{"_76":662,"_452":573,"_453":30,"_166":664},{"_76":673,"_452":570,"_453":30,"_166":250},{"_76":662,"_452":684,"_453":30,"_166":664},{"_76":682,"_452":567,"_453":30,"_166":683},{"_76":679,"_452":680,"_453":30,"_166":681},{"_76":677,"_452":678,"_453":30,"_166":250},{"_76":662,"_452":674,"_453":30,"_166":664},{"_76":676,"_452":567,"_453":30,"_166":250},{"_76":662,"_452":675,"_453":30,"_166":664},{"_76":673,"_452":674,"_453":30,"_166":250},{"_76":672,"_452":670,"_453":84,"_166":250},{"_76":662,"_452":671,"_453":30,"_166":664},{"_76":662,"_452":670,"_453":30,"_166":664},{"_76":662,"_452":669,"_453":30,"_166":664},{"_76":668,"_452":567,"_453":84,"_166":664},{"_76":665,"_452":666,"_453":84,"_166":667},{"_76":662,"_452":663,"_453":30,"_166":664},"_gd_*","info.nvidia.com","This cookie contains no identifying information and is used for diagnostic purposes by Impact Radius, a service we use to track signup commissions for our affiliate program. We have categorized it under Functionality because it is necessary for the operation of our affiliate program.","nv_country_code","developer.nvidia.com","This cookie is likely used to store the country code of the visitor to provide localized content and possibly for legal compliance with regional data privacy laws.","_gdxxxxxxxxxxxxx","nvidianews.nvidia.com","docs.api.nvidia.com","research.nvidia.com","readme_library","wcmmode","ai-podcast.nvidia.com","resources.nvidia.com","geo_locator","openedx-language-preference","learn.nvidia.com","doxygen_width","raytracing-docs.nvidia.com","This cookie is likely used to remember the width preferences of a sidebar or interface element within a documentation platform like Doxygen.","_sfid_eb78","This cookie is categorized as Functionality Cookie.","crowdsource.nvidia.com","mbox","Cookie set by Adobe SiteCatalyst analytics software - used for measuring the performance of page content using A/B split testing","readme_language","vid","This cookie name 'vid' is used for tracking the search request made by the user on the 'Find a practitioner' page.","at_check","Used by Adobe Target to check if cookies are enabled/supported on the browser","blogs.nvidia.com","nvHeadless","This cookie is likely used to track if a user is using the site in a 'headless' non-GUI mode, which could be for accessibility reasons or for automated processes.","C0002","Performance Cookies","These cookies are used to provide quantitative measures of our website visitors, such as the number of times you visit, time on page, your mouse movements, scrolling, clicks and keystroke activity on the websites; other browsing, search, or product research behavior; and what brought you to our site. These cookies may store a unique ID so that our system will remember you when you return. Information collected with these cookies is used to measure and find ways to improve website performance.",[720,721,722,723,724,725,726,727,728,729,730,731,732,733,734,735,736,737,738,739,740,741,742,743,744,745,746,747,748],[700,701,702,703],{"_446":715,"_239":715,"_166":250,"_420":716},{"_446":710,"_239":710,"_166":250,"_420":711},{"_446":508,"_239":508,"_166":250,"_420":707},{"_446":622,"_239":622,"_166":250,"_420":704},[705],{"_76":706,"_452":622,"_453":84,"_166":626},"_gh_sess",[708],{"_76":709,"_452":508,"_453":84,"_166":515},"YSC","bizible.com",[712],{"_76":713,"_452":710,"_453":30,"_166":714},"_BUID","Bizible is a marketing attribution software that tracks the performance of marketing efforts and connects them to revenue, primarily used for B2B environments.","a.usbrowserspeed.com",[717],{"_76":718,"_452":715,"_453":30,"_166":719},"tuid","This domain is owned by Ookla which provides internet speed testing services through its Speedtest application.",{"_76":790,"_452":567,"_453":30,"_166":791},{"_76":775,"_452":789,"_453":84,"_166":776},{"_76":758,"_452":570,"_453":30,"_166":683},{"_76":758,"_452":671,"_453":30,"_166":683},{"_76":787,"_452":567,"_453":30,"_166":788},{"_76":775,"_452":573,"_453":84,"_166":776},{"_76":785,"_452":567,"_453":30,"_166":786},{"_76":775,"_452":570,"_453":84,"_166":776},{"_76":783,"_452":670,"_453":30,"_166":784},{"_76":775,"_452":692,"_453":84,"_166":776},{"_76":755,"_452":570,"_453":30,"_166":757},{"_76":781,"_452":567,"_453":84,"_166":782},{"_76":779,"_452":567,"_453":30,"_166":780},{"_76":777,"_452":567,"_453":84,"_166":778},{"_76":775,"_452":666,"_453":84,"_166":776},{"_76":773,"_452":567,"_453":30,"_166":774},{"_76":771,"_452":567,"_453":30,"_166":772},{"_76":769,"_452":567,"_453":84,"_166":770},{"_76":767,"_452":675,"_453":30,"_166":768},{"_76":755,"_452":666,"_453":30,"_166":757},{"_76":765,"_452":567,"_453":30,"_166":766},{"_76":763,"_452":567,"_453":84,"_166":764},{"_76":761,"_452":567,"_453":30,"_166":762},{"_76":759,"_452":567,"_453":30,"_166":760},{"_76":758,"_452":567,"_453":30,"_166":683},{"_76":755,"_452":756,"_453":30,"_166":757},{"_76":753,"_452":567,"_453":30,"_166":754},{"_76":751,"_452":567,"_453":84,"_166":752},{"_76":749,"_452":567,"_453":30,"_166":750},"_ga_xxxxxxxxxx","Used by Google Analytics to identify and track an individual session with your device.","s_ppvl","This cookie is set by Adobe Analytics to store information on the percentage of the page displayed to the end user.","coveo_visitorId","This cookie name is associated with Coveo which is used to store a unique user ID for marketing purposes and to track the number of distinct users who submitted a query or clicked an item, counting them only once even if they did the same query or clicked the same item over the course of many visits.","_dd_s","marketplace.nvidia.com","Datadog Cookie used to group all events generated from a unique user session across multiple pages. It contains the current session ID, whether the session is excluded due to sampling, and the expiration date of the session. The cookie is extended for an extra 15 minutes every time the user interacts with the website, up to the maximum user session duration (4 hours).","nvweb_A","_ga","This cookie name is associated with Google Universal Analytics - which is a significant update to Google's more commonly used analytics service. This cookie is used to distinguish unique users by assigning a randomly generated number as a client identifier. It is included in each page request in a site and used to calculate visitor, session and campaign data for the sites analytics reports. By default it is set to expire after 2 years, although this is customisable by website owners.","_uetvid","This is a cookie utilised by Microsoft Bing Ads and is a tracking cookie. It allows us to engage with a user that has previously visited our website.","tp","This cookie is associated with Adobe Analytics site tracking. It registers if the website's scroll-depth detection is active. This function remembers how far the user has scrolled on the current session across the website's sub-pages.","_hly_vid","This cookie _hly_vid is used to record website and server performance with each visit to the site. The data is aggregated and analyzed; therefore, it remains anonymous upon use.","_session_id","This cookie creates an interim session ID used as an in-session user ID.\n","s_cc","This cookie is associated with the Adobe Site Catalyst. It determines whether cookies are enabled in the web browser.","s_ecid","Adobe Analytics cookie containing the the Adobe Experience Cloud 1st party visitor id if the Analytics AMCV visitor id is not found.","_hjSessionUser_xxxxxx","Hotjar cookie that is set when a user first lands on a page with the Hotjar script. It is used to persist the Hotjar User ID, unique to that site on the browser. This ensures that behavior in subsequent visits to the same site will be attributed to the same user ID.","_hjHasCachedUserAttributes","This cookie name is associated with Hotjar which is used to know whether the data set in _hjUserAttributes Local Storage item is up to date or not.","s_ppv","This cookie is associated with Adobe Analytics site tracking and is used to measure a user's scroll activity to see how much of a page they view before moving to another page.","_gid","This cookie name is associated with Google Universal Analytics. This appears to be a new cookie and as of Spring 2017 no information is available from Google. It appears to store and update a unique value for each page visited.","_hjTLDTest","When the Hotjar script executes we try to determine the most generic cookie path we should use, instead of the page hostname. This is done so that cookies can be shared across subdomains (where applicable). To determine this, we try to store the _hjTLDTest cookie for different URL substring alternatives until it fails. After this check, the cookie is removed.","ekfls-2025-03-27","This cookie name is associated with readme.io which is used to help maintain data for the API reference support page.","s_getNewRepeat","This cookie is used to determine whether a visitor is a new visitor or a repeat visitor to the site, and is used for analytical purposes.","_hly_sid","This cookie is likely associated with website analytics, as the prefix '_hly_' suggests a connection to Holistics, a business intelligence and analytics platform.","docs.nvidia.com","_hjSession_xxxxxx","A cookie that holds the current session data. This ensues that subsequent requests within the session window will be attributed to the same Hotjar session.","C0001","Required Cookies","These cookies enable core functionality such as security, network management, and accessibility. These cookies are required for the site to function and cannot be turned off.",[937,938,939,940,941,942,943,944,945,946,947,948,949,950,951,952,953,954,955,956,957,958,959,960,961,962,963,964,965,966,967,968,969,970,971,972,973,974,975,976,977,978,979,980,981,982,983,984,985,986,987,988,989,990,991,992,993,994,995,996,997,998,999,1000,1001,1002,1003,1004,1005,1006,1007,1008,1009],[797,798,799,800,801,802,803,804,805,806,807,808,809,810,811,812,813,814,815,816,817,818,819,820,821],{"_446":934,"_239":934,"_166":250,"_420":935},{"_446":928,"_239":928,"_166":250,"_420":929},{"_446":925,"_239":925,"_166":250,"_420":926},{"_446":913,"_239":913,"_166":250,"_420":914},{"_446":910,"_239":910,"_166":250,"_420":911},{"_446":907,"_239":907,"_166":250,"_420":908},{"_446":902,"_239":902,"_166":250,"_420":903},{"_446":508,"_239":508,"_166":250,"_420":899},{"_446":896,"_239":896,"_166":250,"_420":897},{"_446":500,"_239":500,"_166":250,"_420":894},{"_446":495,"_239":495,"_166":250,"_420":892},{"_446":887,"_239":887,"_166":250,"_420":888},{"_446":882,"_239":882,"_166":250,"_420":883},{"_446":877,"_239":877,"_166":250,"_420":878},{"_446":874,"_239":874,"_166":250,"_420":875},{"_446":622,"_239":622,"_166":250,"_420":871},{"_446":480,"_239":480,"_166":250,"_420":868},{"_446":617,"_239":617,"_166":250,"_420":861},{"_446":613,"_239":613,"_166":250,"_420":855},{"_446":849,"_239":849,"_166":250,"_420":850},{"_446":844,"_239":844,"_166":250,"_420":845},{"_446":839,"_239":839,"_166":250,"_420":840},{"_446":835,"_239":835,"_166":250,"_420":836},{"_446":830,"_239":830,"_166":250,"_420":831},{"_446":822,"_239":822,"_166":250,"_420":823},"polyfill.io",[824,825],{"_76":828,"_452":822,"_453":30,"_166":829},{"_76":826,"_452":822,"_453":30,"_166":827},"cf_ob_info","This cookie is associated with sites using CloudFlare. It is used to improve page load times and to override any security restrictions based on the visitors IP address. It does not contain any user identification information.","cf_use_ob","This cookie is associated with sites using CloudFlare. It is used to improve page load times and to override any security restrictions based on the visitors IP address. It does not contain any user identification information. ","www.recaptcha.net",[832],{"_76":833,"_452":830,"_453":30,"_166":834},"_GRECAPTCHA","This domain is used by Google reCAPTCHA, which provides spam protection services to websites by detecting if a user is a human or a bot.","api2.hcaptcha.com",[837],{"_76":838,"_452":835,"_453":30,"_166":489},"__cflb","instagram.com",[841],{"_76":842,"_452":839,"_453":30,"_166":843},"csrftoken","Instagram is a social networking service owned by Facebook, Inc. It's designed for sharing photos and videos from a smartphone. The service also uses cookies for various purposes, including analytics, personalization, and advertising.","readme.io",[846],{"_76":847,"_452":844,"_453":30,"_166":848},"__cf_bm","The __cf_bm cookie is a cookie necessary to support Cloudflare Bot Management, currently in private beta. As part of our bot management service, this cookie helps manage incoming traffic that matches criteria associated with bots.","nvidiacorp.us-5.evergage.com",[851,852],{"_76":854,"_452":849,"_453":30,"_166":489},{"_76":853,"_452":849,"_453":30,"_166":489},"AWSALBTG","AWSALBTGCORS",[856,857],{"_76":859,"_452":613,"_453":30,"_166":860},{"_76":858,"_452":613,"_453":30,"_166":489},"ak_bmsc","bm_mi","This domain is likely associated with NVIDIA's partner program, which could involve marketing, sales, and support services for partners.",[862,863,864],{"_76":867,"_452":617,"_453":30,"_166":489},{"_76":866,"_452":617,"_453":84,"_166":489},{"_76":865,"_452":617,"_453":30,"_166":489},"AWSALBCORS","JSESSIONID","AWSALB",[869],{"_76":847,"_452":480,"_453":30,"_166":870},"This is a CloudFoundry cookie",[872],{"_76":873,"_452":622,"_453":30,"_166":626},"logged_in","hcaptcha.com",[876],{"_76":847,"_452":874,"_453":30,"_166":870},"tracking.intentsify.io",[879],{"_76":880,"_452":877,"_453":30,"_166":881},"userId","Used to track usage of the embedded video player.","static.reo.dev",[884],{"_76":885,"_452":882,"_453":30,"_166":886},"cookietest","Common cookie name could have a number of different origins. Where this is first party and a session cookie, its most likely to do with checking to see if the browser is set to block or allow cookies.","www.datadoghq-browser-agent.com",[889],{"_76":890,"_452":887,"_453":30,"_166":891},"dd_cookie_test_","dd_cookie_test",[893],{"_76":847,"_452":495,"_453":30,"_166":870},[895],{"_76":847,"_452":500,"_453":30,"_166":870},"supabase.co",[898],{"_76":847,"_452":896,"_453":30,"_166":870},[900],{"_76":901,"_452":508,"_453":30,"_166":515},"VISITOR_PRIVACY_METADATA","dosa-prd.azurewebsites.net",[904],{"_76":905,"_452":902,"_453":30,"_166":906},"access_token","This cookie is used to obtain the permissions for users of the portal to allow for API access in order to retrieve information required to perform portal functions.","w.hcaptcha.com",[909],{"_76":847,"_452":907,"_453":30,"_166":870},"cdn.readme.io",[912],{"_76":847,"_452":910,"_453":30,"_166":870},"queue-it.net",[915,916,917,918,919],{"_76":924,"_452":913,"_453":30,"_166":250},{"_76":923,"_452":913,"_453":30,"_166":250},{"_76":922,"_452":913,"_453":30,"_166":250},{"_76":921,"_452":913,"_453":30,"_166":250},{"_76":920,"_452":913,"_453":30,"_166":250},"Queue-it-{queueId}","Queue-it-{dmonvidia }________{eventId}","Queue-it-token","queueitsoftblock_{queueId}","Queue-it-{nvidia}________{eventId}","x.com",[927],{"_76":847,"_452":925,"_453":30,"_166":848},"api.hcaptcha.com",[930,931],{"_76":933,"_452":928,"_453":30,"_166":489},{"_76":838,"_452":928,"_453":30,"_166":932},"This domain is associated with hCaptcha, a service that provides CAPTCHA and reCAPTCHA alternatives for user verification and bot protection.","hmt_id","vimeo.com",[936],{"_76":847,"_452":934,"_453":30,"_166":870},{"_76":1070,"_452":567,"_453":30,"_166":1101},{"_76":1099,"_452":663,"_453":84,"_166":1100},{"_76":1092,"_452":1061,"_453":84,"_166":1093},{"_76":1098,"_452":756,"_453":84,"_166":250},{"_76":1096,"_452":567,"_453":30,"_166":1097},{"_76":854,"_452":1082,"_453":30,"_166":1021},{"_76":1029,"_452":1091,"_453":84,"_166":1031},{"_76":853,"_452":1020,"_453":30,"_166":1021},{"_76":1032,"_452":1091,"_453":30,"_166":1033},{"_76":1094,"_452":567,"_453":84,"_166":1095},{"_76":1092,"_452":1068,"_453":84,"_166":1093},{"_76":1026,"_452":1091,"_453":30,"_166":1028},{"_76":1089,"_452":567,"_453":30,"_166":1090},{"_76":853,"_452":1012,"_453":30,"_166":1021},{"_76":1017,"_452":1088,"_453":30,"_166":1019},{"_76":865,"_452":1018,"_453":30,"_166":1034},{"_76":867,"_452":1018,"_453":30,"_166":1085},{"_76":1032,"_452":1027,"_453":30,"_166":1033},{"_76":1086,"_452":666,"_453":30,"_166":1087},{"_76":867,"_452":692,"_453":30,"_166":1085},{"_76":1083,"_452":567,"_453":30,"_166":1084},{"_76":859,"_452":692,"_453":30,"_166":1078},{"_76":1055,"_452":671,"_453":84,"_166":250},{"_76":853,"_452":1082,"_453":30,"_166":1021},{"_76":1080,"_452":1020,"_453":84,"_166":1081},{"_76":1044,"_452":567,"_453":30,"_166":1035},{"_76":1017,"_452":567,"_453":30,"_166":1019},{"_76":1079,"_452":579,"_453":30,"_166":250},{"_76":859,"_452":567,"_453":30,"_166":1078},{"_76":867,"_452":1012,"_453":30,"_166":1077},{"_76":1075,"_452":675,"_453":84,"_166":1076},{"_76":1073,"_452":567,"_453":84,"_166":1074},{"_76":1072,"_452":567,"_453":30,"_166":250},{"_76":1070,"_452":1018,"_453":30,"_166":1071},{"_76":1069,"_452":579,"_453":30,"_166":250},{"_76":1060,"_452":1068,"_453":84,"_166":1062},{"_76":1066,"_452":671,"_453":30,"_166":1067},{"_76":1064,"_452":756,"_453":30,"_166":1065},{"_76":1063,"_452":573,"_453":30,"_166":250},{"_76":1060,"_452":1061,"_453":84,"_166":1062},{"_76":1058,"_452":1027,"_453":84,"_166":1059},{"_76":1056,"_452":567,"_453":30,"_166":1057},{"_76":1055,"_452":570,"_453":84,"_166":250},{"_76":1026,"_452":1030,"_453":30,"_166":1028},{"_76":858,"_452":1045,"_453":30,"_166":1035},{"_76":1053,"_452":570,"_453":30,"_166":1054},{"_76":1051,"_452":567,"_453":30,"_166":1052},{"_76":1049,"_452":567,"_453":30,"_166":1050},{"_76":1048,"_452":567,"_453":30,"_166":250},{"_76":890,"_452":570,"_453":30,"_166":1047},{"_76":1044,"_452":692,"_453":30,"_166":1035},{"_76":847,"_452":663,"_453":30,"_166":1046},{"_76":1044,"_452":1045,"_453":30,"_166":1035},{"_76":1043,"_452":579,"_453":30,"_166":250},{"_76":1042,"_452":567,"_453":30,"_166":1015},{"_76":1040,"_452":1027,"_453":30,"_166":1041},{"_76":853,"_452":1038,"_453":30,"_166":1039},{"_76":1037,"_452":579,"_453":30,"_166":250},{"_76":1036,"_452":567,"_453":30,"_166":250},{"_76":1029,"_452":1027,"_453":84,"_166":1031},{"_76":858,"_452":692,"_453":30,"_166":1035},{"_76":865,"_452":692,"_453":30,"_166":1034},{"_76":1032,"_452":1030,"_453":30,"_166":1033},{"_76":1029,"_452":1030,"_453":84,"_166":1031},{"_76":1026,"_452":1027,"_453":30,"_166":1028},{"_76":1024,"_452":1020,"_453":84,"_166":1025},{"_76":1022,"_452":678,"_453":30,"_166":1023},{"_76":854,"_452":1020,"_453":30,"_166":1021},{"_76":1017,"_452":1018,"_453":30,"_166":1019},{"_76":847,"_452":670,"_453":30,"_166":1016},{"_76":1014,"_452":567,"_453":30,"_166":1015},{"_76":865,"_452":1012,"_453":30,"_166":1013},{"_76":1010,"_452":567,"_453":30,"_166":1011},"LoginHints","Keeps users logged in across NVIDIA domains.","api.ngc.nvidia.com","This cookie is managed by AWS and is used for load balancing.","bm_so","This cookie name is used to load the site as quickly as possible and to recognize bots using Akamai.","The __cf_bm cookie is a cookie necessary to support Cloudflare Bot Management, currently in private beta. As part of our bot management service, this cookie helps manage incoming traffic that matches criteria associated with bots.\nThis is a CloudFoundry cookie","OptanonConsent","nvid.nvidia.com","This cookie is set by the cookie compliance solution from OneTrust. It stores information about the categories of cookies the site uses and whether visitors have given or withdrawn consent for the use of each category. This enables site owners to prevent cookies in each category from being set in the users browser, when consent is not given. The cookie has a normal lifespan of one year, so that returning visitors to the site will have their preferences remembered. It contains no information that can identify the site visitor.","events.nvidia.com","This cookie is associated with the Amazon Web Services Elastic Load Balancing functionality for routing client request on the server.Used to attribute commission to affiliates when you arrive at the website from an affiliate referral link. It is set when you click on one of our links and used to let the advertiser and us know the website from which you came.","sessionid","This is a very generic cookie name that may have different purposes on different sites, but generally it will be some kind of anonymous session identifier.","_csrf","This cookie is used to prevent Cross-Site Request Forgery (CSRF) attacks, ensuring that each request made to the server is from the user's browser and is intentional.","LSKey-c$CookieConsentPolicy","enterprise-support.nvidia.com","Used to apply end-user cookie consent preferences set by our client-side utility.Custom- Short Duration. Unique ID of an Enquiry- used to validate access to case, and load exisiting data.\n","renderCtx","academicgrants.nvidia.com","Used to store site parameters in the session for reuse across requests by a single client for functionality and performance reasons.","CookieConsentPolicy","Used to apply end-user cookie consent preferences set by our client-side utility.","This cookie is managed by AWS and is used for load balancing.Used to attribute commission to affiliates when you arrive at the website from an affiliate referral link. It is set when you click on one of our links and used to let the advertiser and us know the website from which you came.","This cookie is associated with Akamai and is used to differentiate between traffic from humans and bots.","nv_co","__uxq412__ses.88c1","learn.learn.nvidia.com","This cookie is associated with the Amazon Web Services Elastic Load Balancing functionality for routing client request on the server.","pctrk","Used to count page views by unauthenticated users against license usage.","bm_s","bpazaws52gukakzc__ctrl0_ctl36_uccaptcha","bm_sv","store.nvidia.com","The __cf_bm cookie is a cookie necessary to support Cloudflare Bot Management, currently in private beta. As part of our bot management service, this cookie helps manage incoming traffic that matches criteria associated with bots.\n","This cookie name is used by DataDog to determine if cookies are supported on the website.dd_cookie_test","AMP_TLDTEST_xxxxxxxx","SESSlsd","This cookie is likely to be a session identifier used to maintain user session state across page requests.","mboxEdgeCluster","This cookie name is associated with Adobe Target which is used to communicate the correct edge cluster to at.js so that Target profiles can stay in sync as users navigate across a site.","_rdt_em","This cookie name is associated with Reddit which is used to manage session state and user interactions.","megaNavCollapsed","_abck","The _abck cookie is associated with Akamai's Bot Manager, which is used to prevent fraud and abuse by safeguarding online interactions and detecting irregularities that could suggest malicious activity, such as bot-driven attacks or hacking attempts.","idccsrf","This cookie name is associated with Salesforce which is used to track CrossSiteRequestForgery validation for certain single sign-on (SSO) flows.","ApplicationGatewayAffinity","partners.nvidia.com","This cookie is used to maintain a user session by the server in a load balanced environment, to ensure the user has a consistent experience.","AMP_TEST_xxxxxxxx","bm_lso","This cookie name is associated with Akamai which is used to load the site as quickly as possible and to recognize bots.","SSESS###########","This cookie is likely to be a session cookie used to maintain an anonymized user session by the server in a secure environment, often seen in sites built with the Drupal content management system.","nvdeveloper.nvidia.com","__uxq412__id.88c1","OptanonAlertBoxClosed","This cookie is set by websites using certain versions of the cookie law compliance solution from OneTrust. It is set after visitors have seen a cookie information notice and in some cases only when they actively close the notice down. It enables the website not to show the message more than once to a user. The cookie has a normal lifespan of one year and contains no personal information.","QueueITAccepted-SDFrts345E-V3_{eventId}","TEST_AMCV_COOKIE_WRITE","This cookie is used to test whether cookies are enabled in the browser. It is a temporary cookie used to determine if the browser accepts cookies.","XSRF-TOKEN","This cookie is written to help with site security in preventing Cross-Site Request Forgery attacks.This cookie helps with site security in preventing Cross-Site Request Forgery attacks.","AWS ELB application load balancer","This cookie is associated with Akamai. It is used to analyse traffic to determine if it is automated traffic generated by bots or a human user.","bpazaws52gukakzc__ctrl0_ctl78_list_ctl09_systemcaptcha","PHPSESSID","PHP session cookie associated with embedded content from this domain.\n","base.manager.learn.nvidia.com","bm_ss","This cookie is used to help protect websites from automated bot traffic.","AWS ELB application load balancerUsed to attribute commission to affiliates when you arrive at the website from an affiliate referral link. It is set when you click on one of our links and used to let the advertiser and us know the website from which you came.","_devzone_session","This cookie is likely used to maintain a user session on the website by the server. It is common for session identifiers to be used to keep users logged in and track their interaction with the site during a session.","docs.ngc.nvidia.com","bm_sz","This cookie is used by Akamai to optimize site performance and security. It is typically associated with the delivery of web content and services.","vcalliance.nvidia.com","ASP.NET_SessionId","General purpose platform session cookie, used by sites written with Microsoft .NET based technologies. Usually used to maintain an anonymised user session by the server.","com.adobe.alloy.getTld","This cookie name is associated with the Adobe WebSDK which is used to determine the top-most domain that can accept cookies.","SESSls","This cookie is likely to be a session identifier, used to maintain a user's session state across page requests.","c_code","BIGipServerx","This cookie name is associated with the BIG-IP product suite from company F5. Usually associated with managing sessions on load balanced servers, to ensure user requests are routed consistently to the correct server. The common root is BIGipServer most commonly followed by a domain name, usually the one that it is hosted on, but not always.","This cookie is set by websites using certain versions of the cookie law compliance solution from OneTrust. It is set after visitors have seen a cookie information notice and in some cases only when they actively close the notice down. It enables the website not to show the message more than once to a user. The cookie has a one year lifespan and contains no personal information.","default","

NVIDIA uses cookies to improve your experience on our web site. We and our third-party partners also use cookies and other tools to collect and record information you provide as well as information about your interactions with our websites for performance improvement, analytics, and to assist in marketing efforts. By clicking \"Accept All\", you consent to our use of cookies and other tools as described in our Cookie Policy. You can manage your cookie settings by clicking on \"Manage Settings.\" By continuing to use this site or by clicking one of the buttons below, you agree to our Terms of Service (which contains important waivers). Please see our Privacy Policy for more information on our privacy practices.

","gpc","

We have detected the Global Privacy Control (GPC) signal and recorded your rejection of all optional cookies on this site for this browser. You can manage your cookie settings by clicking on \"Manage Settings\". Please see our Cookie Policy for more information. To opt out of non-cookie personal information \"sales\" / \"sharing\" for targeted advertising purposes, please visit the NVIDIA Preference Center. Please see our Privacy Policy for more information on our privacy practices.

","gpcLoggedIn","

We have detected the Global Privacy Control Signal (GPC) and have opted you out of all optional cookies on this browser. You can manage your cookie settings by clicking on \"Manage Settings\". Please see our Cookie Policy for more information. We have also opted you out of \"sharing\"/\"sales\" of personal information outside of cookies. You can manage these settings in the NVIDIA NVIDIA Preference Center. Please see our Privacy Policy for more information.

","gpcLoggedInOverridden","

We have detected the Global Privacy Control Signal (GPC) and have opted you out of all optional cookies on this browser. You can manage your cookie settings by clicking on \"Manage Settings\". Please see our Cookie Policy for more information. We have also opted you out of \"sharing\"/\"sales\" of personal information outside of cookies which overrides at least one of your previous settings. You can manage them in the NVIDIA Preference Center. Please see our Privacy Policy for more information.

",{"_792":16,"_695":55,"_602":55,"_424":55,"_414":55},"hasConsented","gpcApplied","showBanner","geolocation",{"_1116":1117,"_1118":1119},"country","US","state","OR","query","categories",[1138,1139,1140,1141,1142],"optionUrls",[1126,1127,1128,1129,1130,1131,1132,1133,1134,1135,1136,1137],"total","/orgs/nim/deepseek-ai/containers/deepseek-r1/-","/orgs/nim/nvidia/containers/llama-3.1-nemotron-70b-instruct/-","/orgs/nvidia/-/containers/pytorch/-","/orgs/nvidia/omniverse/collections/kit/-","/orgs/nvidia/-/collections/deepstream_sdk/-","/orgs/nvidia/omniverse/collections/kit-appstreaming-collection/-","/orgs/nvidia/research/models/stylegan3/-","/orgs/nvidia/tao/models/peoplenet/-","/orgs/nvidia/tao/models/trafficcamnet/-","/orgs/nvidia/riva/resources/riva_quickstart/-","/orgs/nvidia/-/helm-charts/gpu-operator/-","/orgs/nvidia/-/helm-charts/k8s-nim-operator/-",{"_232":1215,"_1144":1216,"_1146":55},{"_232":1180,"_1144":1181,"_1146":53},{"_232":233,"_1144":1164,"_1146":1165},{"_232":1156,"_1144":1157,"_1146":1158},{"_232":1143,"_1144":1145,"_1146":1147},"helm-charts","items",[1148,1149],"startIndex",10,{"_232":1143,"_239":1153,"_160":163,"_162":-7,"_234":1154,"_242":243,"_274":1155,"_324":1136},{"_232":1143,"_239":1150,"_160":163,"_162":-7,"_234":1151,"_242":243,"_274":1152,"_324":1137},"NVIDIA NIM Operator","k8s-nim-operator",[],"GPU Operator","gpu-operator",[],"resources",[1159],9,{"_232":1156,"_239":1160,"_160":163,"_162":1161,"_234":1162,"_242":243,"_274":1163,"_324":1135},"Riva Skills Quick Start","riva","riva_quickstart",[],[1166,1167,1168],6,{"_232":233,"_239":1176,"_160":163,"_162":1177,"_234":1178,"_242":243,"_274":1179,"_324":1132},{"_232":233,"_239":1173,"_160":163,"_162":1170,"_234":1174,"_242":243,"_274":1175,"_324":1133},{"_232":233,"_239":1169,"_160":163,"_162":1170,"_234":1171,"_242":243,"_274":1172,"_324":1134},"TrafficCamNet","tao","trafficcamnet",[],"PeopleNet","peoplenet",[],"StyleGAN3 pretrained models","research","stylegan3",[],"collections",[1182,1183,1184],{"_232":1180,"_239":1212,"_160":163,"_162":1186,"_234":1213,"_242":243,"_274":1214,"_324":1129},{"_232":1180,"_239":1209,"_160":163,"_162":-7,"_234":1210,"_242":243,"_274":1211,"_324":1130},{"_232":1180,"_239":1185,"_160":163,"_162":1186,"_234":1187,"_242":243,"_274":1188,"_324":1131},"Omniverse Kit App Streaming","omniverse","kit-appstreaming-collection",[1189,1190,1191],{"_76":1203,"_239":1204,"_236":1205,"_247":1206,"_1196":1207},{"_76":1186,"_239":1199,"_236":1200,"_247":1195,"_1196":1201},{"_76":1192,"_239":1193,"_236":1194,"_247":1195,"_1196":1197},"omniverse-dev","NVIDIA Developer Program","Join the Developer Program for access to free tools, support, and tech resources.","https://store-images.s-microsoft.com/image/apps.33936.6f1bcf42-55d0-45d1-b4b6-f08c68a50bbe.e91ca707-db64-49d6-9761-4781590a9685.7ba64fc5-7c44-4864-a5e6-7bde05126935","supportedFeatures",[1198],"UNIFIED_CATALOG","NVIDIA Omniverse","NVIDIA Omniverse is a platform of APIs, services, and software development kits (SDKs) that enable developers to build generative AI-enabled tools and applications.",[1198,1202],"SHAREABLE_TO_USER_GROUPS","nv-ai-enterprise","NVIDIA AI Enterprise","Accelerate your AI agent development","https://assets.ngc.nvidia.com/products/catalog/images/product-NVAIE.jpg",[1198,1202,1208],"NGC_CONSOLE","DeepStream SDK","deepstream_sdk",[],"Omniverse Kit (FB)","kit",[1189,1190,1191],"containers",[1217,1218,1219],{"_232":1215,"_239":1230,"_160":161,"_162":1231,"_234":1232,"_242":243,"_274":1233,"_324":1126},{"_232":1215,"_239":1223,"_160":161,"_162":163,"_234":1224,"_242":243,"_274":1225,"_324":1127},{"_232":1215,"_239":1220,"_160":163,"_162":-7,"_234":1221,"_242":243,"_274":1222,"_324":1128},"PyTorch","pytorch",[],"Llama-3.1-Nemotron-70B-Instruct","llama-3.1-nemotron-70b-instruct",[1189,1226],{"_76":1227,"_239":1193,"_236":1194,"_247":1228,"_1196":1229},"nim-dev","https://assets.ngc.nvidia.com/products/catalog/images/product-nim-dev.jpg",[1198],"DeepSeek-R1","deepseek-ai","deepseek-r1",[1189,1226],"ownerOrgs",[],"memberOrgs",[],"externalOrgs",[],"apiKey","2b44f86fea50e90ffd2db29ee2028419","consentGranted","applicationId","4c857e4f-5ffa-45fa-b75e-af773c22f0ad","clientToken","pubb6927db33dae60ba5031a186cde0242a","site","datadoghq.com","service","catalog-ui","env","sessionSampleRate",100,"sessionReplaySampleRate","compressIntakeRequests","storeContextsAcrossPages","defaultPrivacyLevel","mask-user-input","trackViewsManually","trackingConsent","not-granted",{"_1263":1264,"_1265":360,"_1266":30,"_1267":1268,"_1144":1269,"_1270":1271,"_1272":1264,"_1273":30,"_1274":30,"_1275":30,"_1276":1277},"appName","ngc-catalog","appOrigin","hideOnPrivateOrg","icon","ObjectsModules2",[1278,1279,1280,1281,1282,1283,1284],"menuIcon","editor-swatches","productName","requiresAllScopes","requiresLogin","requiresOrgLevelScopesOnly","title","Catalog",{"_1276":1306,"_46":1307,"_1270":1308},{"_1299":1300,"_1276":1301,"_46":1302,"_1272":1203},{"_1276":1297,"_46":1298,"_1270":1271},{"_1276":1294,"_46":1295,"_1270":1296},{"_1276":1291,"_46":1292,"_1270":1293},{"_1276":1288,"_46":1289,"_1270":1290},{"_1276":1285,"_46":1286,"_1270":1287},"Resources","/resources","misc-library","Models","/models","misc-neural-network","Helm Charts","/helm-charts","misc-ship-wheel","Containers","/containers","3d-cube","Collections","/collections","access",[1303],"Enterprise Catalog","/enterprise",[1304,1305],"NVAIE","READ","Explore Catalog","/","common-home",{"_110":111,"_112":113,"_114":115,"_116":117,"_118":119,"_120":121},{"_109":30}]