[{"_1":2,"_3":-5,"_4":-5},"loaderData",{"_5":6,"_7":8,"_9":10,"_11":12},"actionData","errors","root",{"_532":533,"_534":535,"_536":537,"_538":539,"_540":541,"_542":20,"_543":544,"_545":546,"_547":20,"_548":20,"_549":20,"_550":20,"_551":552,"_553":-5,"_554":555,"_556":557,"_558":559,"_560":561,"_562":563,"_564":565,"_566":-5,"_567":20,"_568":-5,"_569":570,"_571":572,"_80":573,"_574":575,"_576":577,"_578":579,"_580":581,"_582":20,"_15":16,"_17":583,"_19":20,"_21":584,"_23":585,"_25":586},"routes/orgs/$orgName/$teamName/$resourceType/$resourceId/$version/route",{"_65":66,"_13":67,"_68":69,"_70":71,"_72":73,"_74":75,"_76":20,"_77":20,"_78":79,"_80":81,"_82":-5,"_83":84,"_85":-5,"_86":87,"_88":89,"_90":20,"_91":92,"_15":16,"_17":93,"_19":20,"_21":94,"_23":95,"_25":96},"routes/orgs/$orgName/$teamName/$resourceType/$resourceId/$version/model-card/route",{"_15":16,"_17":58,"_19":20,"_21":59,"_23":60,"_25":61},"routes/orgs/$orgName/$teamName/$resourceType/$resourceId/$version/model-card/bias/route",{"_13":14,"_15":16,"_17":18,"_19":20,"_21":22,"_23":24,"_25":26},"html","

Bias

\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\">

Field:	Response:
Participation considerations from adversely impacted groups (protected classes) in model design and testing:	None
Measures taken to mitigate against unwanted bias:	None

","traceId","0896c438073677b9cfba2580191ab1a3","traceStartTime",1787766360730,"isStaleResponse",false,"notifications",[],"toasts",[],"apiCalls",[27,28,29,30,31,32],{"_33":48,"_35":56,"_37":50,"_39":51,"_41":57,"_43":44},{"_33":48,"_35":54,"_37":50,"_39":51,"_41":55,"_43":44},{"_33":48,"_35":53,"_37":50,"_39":51,"_41":52,"_43":44},{"_33":48,"_35":49,"_37":50,"_39":51,"_41":52,"_43":44},{"_33":45,"_35":46,"_37":38,"_39":40,"_41":47,"_43":44},{"_33":34,"_35":36,"_37":38,"_39":40,"_41":42,"_43":44},"apiTag","subscription-api","path","/v2/users/me/subscriptions","status","error","httpStatus",401,"durationMs",3,"retryCount",0,"oms-api","/v2/users/me",4,"models-api","/v2/models/org/nim/team/nvidia/llama-3.3-nemotron-super-49b-v1.5/collections?page-number=0&page-size=1","success",200,62,"/v2/models/nim/nvidia/llama-3.3-nemotron-super-49b-v1.5/versions/h200-nvlx2-latency-fp8-as64uhv7eg","/v2/models/nim/nvidia/llama-3.3-nemotron-super-49b-v1.5/versions?page-size=100&page-number=0&sort-order=CREATED_DATE_DESC",79,"/v2/models/nim/nvidia/llama-3.3-nemotron-super-49b-v1.5?resolve-labels=true&remove-unresolved-labels=true",157,1787766360556,[],[],[62,63],{"_33":45,"_35":46,"_37":64,"_39":-5,"_41":-5,"_43":44},{"_33":34,"_35":36,"_37":64,"_39":-5,"_41":-5,"_43":44},"pending","resource",{"_236":439,"_440":441,"_442":443,"_444":445,"_103":104,"_105":106,"_446":447,"_109":448,"_449":450,"_114":451,"_452":453,"_454":435,"_455":-7,"_456":-7,"_457":458,"_459":460,"_461":73,"_462":463,"_464":465,"_466":-7,"_467":-7,"_468":469,"_470":471,"_472":473,"_474":475,"_476":477,"_478":479,"_480":481,"_482":483,"_484":485,"_486":20},"

Llama-3.3-Nemotron-Super-49B-v1.5 Overview

Description:

Llama-3.3-Nemotron-Super-49B-v1.5 is a significantly upgraded version of Llama-3.3-Nemotron-Super-49B-v1 and is a large language model (LLM) which is a derivative of Meta Llama-3.3-70B-Instruct (AKA the reference model). It is a reasoning model that is post trained for reasoning, human chat preferences, and agentic tasks, such as Retrieval-Augmented Generation (RAG) and tool calling. The model supports a context length of 128K tokens.

Llama-3.3-Nemotron-Super-49B-v1.5 is a model which offers a great tradeoff between model accuracy and efficiency. Efficiency (throughput) directly translates to savings. Using a novel Neural Architecture Search (NAS) approach, we greatly reduce the model’s memory footprint, enabling larger workloads, as well as fitting the model on a single GPU at high workloads (H200). This NAS approach enables the selection of a desired point in the accuracy-efficiency tradeoff. For more information on the NAS approach, please refer to this paper.

The model underwent a multi-phase post-training process to enhance both its reasoning and non-reasoning capabilities. This includes a supervised fine-tuning stage for Math, Code, Science, and Tool Calling. Additionally, the model went through multiple stages of Reinforcement Learning (RL) including Reward-aware Preference Optimization (RPO) for chat, Reinforcement Learning with Verifiable Rewards (RLVR) for reasoning, and iterative Direct Preference Optimization (DPO) for Tool Calling capability enhancements. The final checkpoint was achieved after merging several RL and DPO checkpoints.

This model is part of the Llama Nemotron Collection. You can find the other model(s) in this family here:

This model is ready for commercial use.

License/Terms of Use

GOVERNING TERMS: The NIM container is governed by the NVIDIA Software License Agreement and the Product-Specific Terms for NVIDIA AI Products; and the use of this model is governed by the NVIDIA Open Model License.

Additional Information: Llama 3.3 Community License Agreement. Built with Llama.

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.\nYou are responsible for ensuring that your use of NVIDIA provided models complies with all applicable laws.

Deployment Geography:

Global

Use Case:

Llama-3.3-Nemotron-Super-49B-v1.5 is a general purpose reasoning and chat model intended to be used in English and coding languages. Other non-English languages (German, French, Italian, Portuguese, Hindi, Spanish, and Thai) are also supported. Developers designing AI Agent systems, chatbots, RAG systems, and other AI-powered applications. Also suitable for typical instruction-following tasks.

Release Date:

Hugging Face 7/25/2025 via Llama-3_3-Nemotron-Super-49B-v1_5
Build.NVIDIA.com 7/25/2025 Llama-3_3-Nemotron-Super-49B-v1_5

References

Model Architecture

Architecture Type: Dense decoder-only Transformer model

Network Architecture: Llama 3.3 70B Instruct, customized through Neural Architecture Search (NAS)

The model is a derivative of Meta’s Llama-3.3-70B-Instruct, using Neural Architecture Search (NAS). The NAS algorithm results in non-standard and non-repetitive blocks. This includes the following:

Skip attention: In some blocks, the attention is skipped entirely, or replaced with a single linear layer. Variable FFN: The expansion/compression ratio in the FFN layer is different between blocks.

We utilize a block-wise distillation of the reference model, where for each block we create multiple variants providing different tradeoffs of quality vs. computational complexity, discussed in more depth below. We then search over the blocks to create a model which meets the required throughput and memory (optimized for a single H100-80GB GPU) while minimizing the quality degradation. The model then undergoes knowledge distillation (KD), with a focus on English single and multi-turn chat use cases. The KD step included 40 billion tokens consisting of a mixture of 3 datasets - FineWeb, Buzz-V1.2, and Dolma.

Intended Use
\nLlama-3.3-Nemotron-Super-49B-v1.5 is a general purpose reasoning and chat model intended to be used in English and coding languages. Other non-English languages (German, French, Italian, Portuguese, Hindi, Spanish, and Thai) are also supported.

Input

Input Type(s): Text
Input Format: String
Input Parameters: One-Dimensional (1D)
Other Properties Related to Input: Context length up to 131,072 tokens

Output

Output Type(s): Text
Output Format: String
Output Parameters: One-Dimensional (1D)
Other Properties Related to Output: Context length up to 131,072 tokens

Our AI models are designed and/or 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: vLLM, TensorRT
Supported Hardware Microarchitecture Compatibility:\n
- NVIDIA Ampere
- NVIDIA Hopper
\n
Preferred 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:

Llama-3.3-Nemotron-Super-49B-v1.5-1.12.0
\nLlama-3.3-Nemotron-Super-49B-v1.5-1.13.1
\nLlama-3.3-Nemotron-Super-49B-v1.5-1.14.0
\nLlama-3.3-Nemotron-Super-49B-v1.5-1.15.0
\nLlama-3.3-Nemotron-Super-49B-v1.5-1.15.5

Quick Start and Usage Recommendations:

By default (empty system prompt), the model will respond in reasoning ON mode. Setting /no_think in the system prompt will enable reasoning OFF mode.
We recommend setting temperature to 0.6 and Top P to 0.95 for Reasoning ON mode.
We recommend using greedy decoding for Reasoning OFF mode.

You can try this model out through the preview API, using this link: Llama-3_3-Nemotron-Super-49B-v1_5.

Use It with vLLM

pip install vllm==0.9.2

An example on how to serve with vLLM:

$ python3 -m vllm.entrypoints.openai.api_server \\\n  --model \"nvidia/Llama-3_3-Nemotron-Super-49B-v1_5\" \\\n  --trust-remote-code \\\n  --seed=1 \\\n  --host=\"0.0.0.0\" \\\n  --port=5000 \\\n  --served-model-name \"Llama-3_3-Nemotron-Super-49B-v1_5\" \\\n  --tensor-parallel-size=8 \\\n  --max-model-len=65536 \\\n  --gpu-memory-utilization 0.95 \\\n  --enforce-eager\n

Running a vLLM Server with Tool-call Support

To enable tool calling usage with this model, we provide a tool parser in the repository. Here is an example on how to use it:

$ git clone https://huggingface.co/nvidia/Llama-3_3-Nemotron-Super-49B-v1_5\n\n$ conda create -n vllm python=3.12 -y\n$ conda activate vllm\n$ pip install vllm==0.9.2\n\n$ python3 -m vllm.entrypoints.openai.api_server \\\n  --model Llama-3_3-Nemotron-Super-49B-v1_5 \\\n  --trust-remote-code \\\n  --seed=1 \\\n  --host=\"0.0.0.0\" \\\n  --port=5000 \\\n  --served-model-name \"Llama-3_3-Nemotron-Super-49B-v1_5\" \\\n  --tensor-parallel-size=8 \\\n  --max-model-len=65536 \\\n  --gpu-memory-utilization 0.95 \\\n  --enforce-eager \\\n  --enable-auto-tool-choice \\\n  --tool-parser-plugin \"Llama-3_3-Nemotron-Super-49B-v1_5/llama_nemotron_toolcall_parser_no_streaming.py\" \\\n  --tool-call-parser \"llama_nemotron_json\"\n

After launching a vLLM server, you can call the server with tool-call support using a Python script like below.

\nOkay, let's see. The user has a bill of $100 and wants to know the amount for an 18% tip. Hmm, I need to calculate the tip based on the bill total and the percentage. The tools provided include calculate_tip, which takes bill_total and tip_percentage as parameters. So the bill_total here is 100, and the tip_percentage is 18. I should call the calculate_tip function with these values. Wait, do I need to check if the parameters are integers? The bill is $100, which is an integer, and 18% is also an integer. So that fits the function's requirements. I don't need to convert any currency here because the user is asking about a tip in the same currency. So the correct tool to use is calculate_tip with those parameters.\n\n'''\nprint(completion.choices[0].message.tool_calls)\n'''\n[ChatCompletionMessageToolCall(id='chatcmpl-tool-e341c6954d2c48c2a0e9071c7bdefd8b', function=Function(arguments='{"bill_total": 100, "tip_percentage": 18}', name='calculate_tip'), type='function')]\n'''\n\">

from openai import OpenAI\nclient = OpenAI(\n    base_url=\"http://0.0.0.0:5000/v1\",\n    api_key=\"dummy\",\n)\ncompletion = client.chat.completions.create(\n    model=\"Llama-3_3-Nemotron-Super-49B-v1_5\",\n    messages=[\n        {\"role\": \"system\", \"content\": \"\"},\n        {\"role\": \"user\", \"content\": \"My bill is $100. What will be the amount for 18% tip?\"}\n    ],\n    tools=[\n        {\n            \"type\": \"function\",\n            \"function\": {\n                \"name\": \"calculate_tip\",\n                \"parameters\": {\n                    \"type\": \"object\",\n                    \"properties\": {\n                        \"bill_total\": {\n                            \"type\": \"integer\",\n                            \"description\": \"The total amount of the bill\"\n                        },\n                        \"tip_percentage\": {\n                            \"type\": \"integer\",\n                            \"description\": \"The percentage of tip to be applied\"\n                        }\n                    },\n                    \"required\": [\"bill_total\", \"tip_percentage\"]\n                }\n            }\n        },\n        {\n            \"type\": \"function\",\n            \"function\": {\n                \"name\": \"convert_currency\",\n                \"parameters\": {\n                    \"type\": \"object\",\n                    \"properties\": {\n                        \"amount\": {\n                            \"type\": \"integer\",\n                            \"description\": \"The amount to be converted\"\n                        },\n                        \"from_currency\": {\n                            \"type\": \"string\",\n                            \"description\": \"The currency code to convert from\"\n                        },\n                        \"to_currency\": {\n                            \"type\": \"string\",\n                            \"description\": \"The currency code to convert to\"\n                        }\n                    },\n                    \"required\": [\"from_currency\", \"amount\", \"to_currency\"]\n                }\n            }\n        }\n    ],\n    temperature=0.6,\n    top_p=0.95,\n    max_tokens=32768,\n    stream=False\n)\nprint(completion.choices[0].message.content)\n'''\n<think>\nOkay, let's see. The user has a bill of $100 and wants to know the amount for an 18% tip. Hmm, I need to calculate the tip based on the bill total and the percentage. The tools provided include calculate_tip, which takes bill_total and tip_percentage as parameters. So the bill_total here is 100, and the tip_percentage is 18. I should call the calculate_tip function with these values. Wait, do I need to check if the parameters are integers? The bill is $100, which is an integer, and 18% is also an integer. So that fits the function's requirements. I don't need to convert any currency here because the user is asking about a tip in the same currency. So the correct tool to use is calculate_tip with those parameters.\n</think>\n'''\nprint(completion.choices[0].message.tool_calls)\n'''\n[ChatCompletionMessageToolCall(id='chatcmpl-tool-e341c6954d2c48c2a0e9071c7bdefd8b', function=Function(arguments='{\"bill_total\": 100, \"tip_percentage\": 18}', name='calculate_tip'), type='function')]\n'''\n

Training, Testing, and Evaluation Datasets:

Training Dataset:

A large variety of training data was used for the knowledge distillation phase before the post-training pipeline, 3 of which included: FineWeb, Buzz-V1.2, and Dolma.

The data for the multi-stage post-training phases for improvements in Code, Math, and Reasoning is a compilation of SFT and RL data that supports improvements of math, code, general reasoning, and instruction following capabilities of the original Llama instruct model.

Prompts have been sourced from either public and open corpus or synthetically generated. Responses were synthetically generated by a variety of models, with some prompts containing responses for both reasoning on and off modes, to train the model to distinguish between two modes.

NVIDIA will be releasing the post-training dataset in the coming weeks.

Data Modality: Text

Data Collection Method by dataset: Hybrid: Automated, Human, Synthetic

Labeling Method by dataset: Hybrid: Automated, Human, Synthetic

Properties (Quantity, Dataset Descriptions, Sensor(s)):

Quantity: The knowledge distillation phase used 40 billion tokens. The underlying Dolma dataset contains approximately 3 trillion tokens, and FineWeb contains over 15 trillion tokens.

Dataset Descriptions: The training data is a composite of large-scale web text, synthetically generated conversational data for coding and instruction-following, and a diverse corpus of text from academic, literary, and encyclopedic sources. The model supports a context length of 128,000 tokens.

Testing Dataset:

Data Collection Method by dataset: Hybrid: Human, Synthetic

Labeling Method by dataset: Hybrid: Human, Automated

Properties (Quantity, Dataset Descriptions, Sensor(s)):

Quantity: The benchmarks contain thousands of test items, including:

MMLU: ~15,900 multiple-choice questions.
GPQA: 448 difficult multiple-choice questions.
HumanEval: 164 programming problems.
GSM-8K: ~8,500 grade-school math problems.
MATH: 12,500 competition math problems.
IF-Eval: Over 400 instruction-following prompts.

Dataset Descriptions: The evaluation suite is diverse, covering 57 academic and professional subjects (MMLU), expert-level reasoning (GPQA), Python code generation (HumanEval), mathematical problem-solving (GSM-8K, MATH), and the ability to follow precise instructions (IF-Eval).

Evaluation Dataset:

We used the datasets listed below to evaluate Llama-3.3-Nemotron-Super-49B-v1.5.

Data Collection Method by dataset: Hybrid: Human. Synthetic

Labeling Method by dataset: Hybrid: Human, Synthetic, Automatic

Evaluation Results
\nWe evaluate the model using temperature=0.6, top_p=0.95, and 64k sequence length. We run the benchmarks up to 16 times and average the scores to be more accurate.

MATH500

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

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

Reasoning Mode	pass@1 (avg. over 4 runs)
Reasoning On	97.4

AIME 2024

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

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

Reasoning Mode	pass@1 (avg. over 16 runs)
Reasoning On	87.5

AIME 2025

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

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

Reasoning Mode	pass@1 (avg. over 16 runs)
Reasoning On	82.71

GPQA

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

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

Reasoning Mode	pass@1 (avg. over 4 runs)
Reasoning On	71.97

LiveCodeBench 24.10-25.02

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

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

Reasoning Mode	pass@1 (avg. over 4 runs)
Reasoning On	73.58

BFCL v3

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

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

Reasoning Mode	pass@1 (avg. over 2 runs)
Reasoning On	71.75

IFEval

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

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

Reasoning Mode	Strict:Instruction
Reasoning On	88.61

ArenaHard

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

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

Reasoning Mode	pass@1 (avg. over 1 runs)
Reasoning On	92.0

Humanity's Last Exam (Text-Only Subset)

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

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

Reasoning Mode	pass@1 (avg. over 1 runs)
Reasoning On	7.64

MMLU Pro (CoT)

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

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

Reasoning Mode	pass@1 (avg. over 1 runs)
Reasoning On	79.53

All evaluations were done using the NeMo-Skills repository.

Inference:

Acceleration Engine:

vLLM, TensorRT

Test Hardware:

2x NVIDIA H100-80GB
2x NVIDIA A100-80GB GPUs
B200 SXM
H200 SXM
H100 SXM
A100 SXM 80GB
A100 SXM 40GB
L40S PCIe
A10G
H100 NVL
H200 NVL
GH200 96GB
GB200
RTX 5090
RTX 4090
RTX 6000 Ada

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.

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

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

You are responsible for ensuring that your use of NVIDIA provided models complies with all applicable laws.

Citation

@misc{bercovich2025llamanemotronefficientreasoningmodels,\n      title={Llama-Nemotron: Efficient Reasoning Models}, \n      author={Akhiad Bercovich and Itay Levy and Izik Golan and Mohammad Dabbah and Ran El-Yaniv and Omri Puny and Ido Galil and Zach Moshe and Tomer Ronen and Najeeb Nabwani and Ido Shahaf and Oren Tropp and Ehud Karpas and Ran Zilberstein and Jiaqi Zeng and Soumye Singhal and Alexander Bukharin and Yian Zhang and Tugrul Konuk and Gerald Shen and Ameya Sunil Mahabaleshwarkar and Bilal Kartal and Yoshi Suhara and Olivier Delalleau and Zijia Chen and Zhilin Wang and David Mosallanezhad and Adi Renduchintala and Haifeng Qian and Dima Rekesh and Fei Jia and Somshubra Majumdar and Vahid Noroozi and Wasi Uddin Ahmad and Sean Narenthiran and Aleksander Ficek and Mehrzad Samadi and Jocelyn Huang and Siddhartha Jain and Igor Gitman and Ivan Moshkov and Wei Du and Shubham Toshniwal and George Armstrong and Branislav Kisacanin and Matvei Novikov and Daria Gitman and Evelina Bakhturina and Jane Polak Scowcroft and John Kamalu and Dan Su and Kezhi Kong and Markus Kliegl and Rabeeh Karimi and Ying Lin and Sanjeev Satheesh and Jupinder Parmar and Pritam Gundecha and Brandon Norick and Joseph Jennings and Shrimai Prabhumoye and Syeda Nahida Akter and Mostofa Patwary and Abhinav Khattar and Deepak Narayanan and Roger Waleffe and Jimmy Zhang and Bor-Yiing Su and Guyue Huang and Terry Kong and Parth Chadha and Sahil Jain and Christine Harvey and Elad Segal and Jining Huang and Sergey Kashirsky and Robert McQueen and Izzy Putterman and George Lam and Arun Venkatesan and Sherry Wu and Vinh Nguyen and Manoj Kilaru and Andrew Wang and Anna Warno and Abhilash Somasamudramath and Sandip Bhaskar and Maka Dong and Nave Assaf and Shahar Mor and Omer Ullman Argov and Scot Junkin and Oleksandr Romanenko and Pedro Larroy and Monika Katariya and Marco Rovinelli and Viji Balas and Nicholas Edelman and Anahita Bhiwandiwalla and Muthu Subramaniam and Smita Ithape and Karthik Ramamoorthy and Yuting Wu and Suguna Varshini Velury and Omri Almog and Joyjit Daw and Denys Fridman and Erick Galinkin and Michael Evans and Katherine Luna and Leon Derczynski and Nikki Pope and Eileen Long and Seth Schneider and Guillermo Siman and Tomasz Grzegorzek and Pablo Ribalta and Monika Katariya and Joey Conway and Trisha Saar and Ann Guan and Krzysztof Pawelec and Shyamala Prayaga and Oleksii Kuchaiev and Boris Ginsburg and Oluwatobi Olabiyi and Kari Briski and Jonathan Cohen and Bryan Catanzaro and Jonah Alben and Yonatan Geifman and Eric Chung and Chris Alexiuk},\n      year={2025},\n      eprint={2505.00949},\n      archivePrefix={arXiv},\n      primaryClass={cs.CL},\n      url={https://arxiv.org/abs/2505.00949}, \n}\n

","resourceVersions",[135,136,137,138,139,140,141,142,143,144,145,146,147,148,149,150,151,152,153,154,155,156,157,158,159,160,161,162,163,164,165,166,167,168,169,170,171,172,173,174,175,176,177,178,179,180,181,182,183,184,185,186,187,188,189,190,191,192,193,194,195,196,197,198,199,200,201,202,203,204,205,206,207,208,209,210,211,212,213,214,215,216,217,218,219,220,221,222,223,224,225,226,227,228,229,230,231,232,233,234,235],"resourceVersionsTotal",142,"hasMoreVersions",true,"versionMetadata",{"_108":81,"_109":-5,"_37":110,"_111":-5,"_112":113,"_114":115,"_116":117,"_118":119,"_120":73,"_121":-5,"_122":20,"_123":-5,"_124":20,"_125":-5,"_126":20,"_127":-5,"_128":20,"_129":-5,"_130":20,"_131":-5,"_132":20,"_133":134},"versionNotFound","isLoggedIn","latestVersionPath","/orgs/nim/nvidia/models/llama-3.3-nemotron-super-49b-v1.5/-","version","h200-nvlx2-latency-fp8-as64uhv7eg","imageReference","cliDownloadCommand","ngc registry model download-version nim/nvidia/llama-3.3-nemotron-super-49b-v1.5:h200-nvlx2-latency-fp8-as64uhv7eg","helmFetchCommand","viewAllVersionsPath","/orgs/nim/nvidia/models/llama-3.3-nemotron-super-49b-v1.5/h200-nvlx2-latency-fp8-as64uhv7eg/version-history","accessCheckResult",{"_107":73},"hasRelatedCollections","context",{"_103":104,"_105":106},1787766360731,[],[],[97,98,99,100,101,102],{"_33":48,"_35":56,"_37":50,"_39":51,"_41":57,"_43":44},{"_33":48,"_35":54,"_37":50,"_39":51,"_41":55,"_43":44},{"_33":48,"_35":53,"_37":50,"_39":51,"_41":52,"_43":44},{"_33":48,"_35":49,"_37":50,"_39":51,"_41":52,"_43":44},{"_33":45,"_35":46,"_37":38,"_39":40,"_41":47,"_43":44},{"_33":34,"_35":36,"_37":38,"_39":40,"_41":42,"_43":44},"orgName","nim","teamName","nvidia","allowed","versionId","description","UPLOAD_COMPLETE","releaseType","createdDate",["D",1760724151243],"updatedDate",["D",1760724160472],"totalSizeInBytes",52484744371,"totalFileCount",5,"isSigned","malwareScanStatus","isScanned","releaseNotes","hasReleaseNotes","setup","hasSetup","quickStartGuide","hasQuickStartGuide","advanced","hasAdvanced","performance","hasPerformance","policies",[],{"_236":437,"_238":438,"_111":-5,"_120":73,"_122":20,"_133":436,"_124":20},{"_236":435,"_238":435,"_111":-5,"_120":73,"_122":20,"_133":436,"_124":20},{"_236":433,"_238":433,"_111":-5,"_120":73,"_122":20,"_133":434,"_124":20},{"_236":431,"_238":431,"_111":-5,"_120":73,"_122":20,"_133":432,"_124":20},{"_236":429,"_238":429,"_111":-5,"_120":73,"_122":20,"_133":430,"_124":20},{"_236":427,"_238":427,"_111":-5,"_120":73,"_122":20,"_133":428,"_124":20},{"_236":425,"_238":425,"_111":-5,"_120":73,"_122":20,"_133":426,"_124":20},{"_236":423,"_238":423,"_111":-5,"_120":73,"_122":20,"_133":424,"_124":20},{"_236":421,"_238":421,"_111":-5,"_120":73,"_122":20,"_133":422,"_124":20},{"_236":419,"_238":419,"_111":-5,"_120":73,"_122":20,"_133":420,"_124":20},{"_236":417,"_238":417,"_111":-5,"_120":73,"_122":20,"_133":418,"_124":20},{"_236":415,"_238":415,"_111":-5,"_120":73,"_122":20,"_133":416,"_124":20},{"_236":413,"_238":413,"_111":-5,"_120":73,"_122":20,"_133":414,"_124":20},{"_236":411,"_238":411,"_111":-5,"_120":73,"_122":20,"_133":412,"_124":20},{"_236":409,"_238":409,"_111":-5,"_120":73,"_122":20,"_133":410,"_124":20},{"_236":407,"_238":407,"_111":-5,"_120":73,"_122":20,"_133":408,"_124":20},{"_236":405,"_238":405,"_111":-5,"_120":73,"_122":20,"_133":406,"_124":20},{"_236":403,"_238":403,"_111":-5,"_120":73,"_122":20,"_133":404,"_124":20},{"_236":401,"_238":401,"_111":-5,"_120":73,"_122":20,"_133":402,"_124":20},{"_236":399,"_238":399,"_111":-5,"_120":73,"_122":20,"_133":400,"_124":20},{"_236":397,"_238":397,"_111":-5,"_120":73,"_122":20,"_133":398,"_124":20},{"_236":395,"_238":395,"_111":-5,"_120":73,"_122":20,"_133":396,"_124":20},{"_236":393,"_238":393,"_111":-5,"_120":73,"_122":20,"_133":394,"_124":20},{"_236":391,"_238":391,"_111":-5,"_120":73,"_122":20,"_133":392,"_124":20},{"_236":389,"_238":389,"_111":-5,"_120":73,"_122":20,"_133":390,"_124":20},{"_236":387,"_238":387,"_111":-5,"_120":73,"_122":20,"_133":388,"_124":20},{"_236":385,"_238":385,"_111":-5,"_120":73,"_122":20,"_133":386,"_124":20},{"_236":383,"_238":383,"_111":-5,"_120":73,"_122":20,"_133":384,"_124":20},{"_236":381,"_238":381,"_111":-5,"_120":73,"_122":20,"_133":382,"_124":20},{"_236":379,"_238":379,"_111":-5,"_120":73,"_122":20,"_133":380,"_124":20},{"_236":377,"_238":377,"_111":-5,"_120":73,"_122":20,"_133":378,"_124":20},{"_236":375,"_238":375,"_111":-5,"_120":73,"_122":20,"_133":376,"_124":20},{"_236":373,"_238":373,"_111":-5,"_120":73,"_122":20,"_133":374,"_124":20},{"_236":371,"_238":371,"_111":-5,"_120":73,"_122":20,"_133":372,"_124":20},{"_236":369,"_238":369,"_111":-5,"_120":73,"_122":20,"_133":370,"_124":20},{"_236":367,"_238":367,"_111":-5,"_120":73,"_122":20,"_133":368,"_124":20},{"_236":365,"_238":365,"_111":-5,"_120":73,"_122":20,"_133":366,"_124":20},{"_236":363,"_238":363,"_111":-5,"_120":73,"_122":20,"_133":364,"_124":20},{"_236":361,"_238":361,"_111":-5,"_120":73,"_122":20,"_133":362,"_124":20},{"_236":359,"_238":359,"_111":-5,"_120":73,"_122":20,"_133":360,"_124":20},{"_236":357,"_238":357,"_111":-5,"_120":73,"_122":20,"_133":358,"_124":20},{"_236":355,"_238":355,"_111":-5,"_120":73,"_122":20,"_133":356,"_124":20},{"_236":353,"_238":353,"_111":-5,"_120":73,"_122":20,"_133":354,"_124":20},{"_236":351,"_238":351,"_111":-5,"_120":73,"_122":20,"_133":352,"_124":20},{"_236":349,"_238":349,"_111":-5,"_120":73,"_122":20,"_133":350,"_124":20},{"_236":347,"_238":347,"_111":-5,"_120":73,"_122":20,"_133":348,"_124":20},{"_236":345,"_238":345,"_111":-5,"_120":73,"_122":20,"_133":346,"_124":20},{"_236":343,"_238":343,"_111":-5,"_120":73,"_122":20,"_133":344,"_124":20},{"_236":341,"_238":341,"_111":-5,"_120":73,"_122":20,"_133":342,"_124":20},{"_236":339,"_238":339,"_111":-5,"_120":73,"_122":20,"_133":340,"_124":20},{"_236":337,"_238":337,"_111":-5,"_120":73,"_122":20,"_133":338,"_124":20},{"_236":335,"_238":335,"_111":-5,"_120":73,"_122":20,"_133":336,"_124":20},{"_236":333,"_238":333,"_111":-5,"_120":73,"_122":20,"_133":334,"_124":20},{"_236":331,"_238":331,"_111":-5,"_120":73,"_122":20,"_133":332,"_124":20},{"_236":329,"_238":329,"_111":-5,"_120":73,"_122":20,"_133":330,"_124":20},{"_236":327,"_238":327,"_111":-5,"_120":73,"_122":20,"_133":328,"_124":20},{"_236":325,"_238":325,"_111":-5,"_120":73,"_122":20,"_133":326,"_124":20},{"_236":323,"_238":323,"_111":-5,"_120":73,"_122":20,"_133":324,"_124":20},{"_236":321,"_238":321,"_111":-5,"_120":73,"_122":20,"_133":322,"_124":20},{"_236":319,"_238":319,"_111":-5,"_120":73,"_122":20,"_133":320,"_124":20},{"_236":317,"_238":317,"_111":-5,"_120":73,"_122":20,"_133":318,"_124":20},{"_236":315,"_238":315,"_111":-5,"_120":73,"_122":20,"_133":316,"_124":20},{"_236":313,"_238":313,"_111":-5,"_120":73,"_122":20,"_133":314,"_124":20},{"_236":311,"_238":311,"_111":-5,"_120":73,"_122":20,"_133":312,"_124":20},{"_236":309,"_238":309,"_111":-5,"_120":73,"_122":20,"_133":310,"_124":20},{"_236":307,"_238":307,"_111":-5,"_120":73,"_122":20,"_133":308,"_124":20},{"_236":305,"_238":305,"_111":-5,"_120":73,"_122":20,"_133":306,"_124":20},{"_236":303,"_238":303,"_111":-5,"_120":73,"_122":20,"_133":304,"_124":20},{"_236":301,"_238":301,"_111":-5,"_120":73,"_122":20,"_133":302,"_124":20},{"_236":299,"_238":299,"_111":-5,"_120":73,"_122":20,"_133":300,"_124":20},{"_236":297,"_238":297,"_111":-5,"_120":73,"_122":20,"_133":298,"_124":20},{"_236":295,"_238":295,"_111":-5,"_120":73,"_122":20,"_133":296,"_124":20},{"_236":81,"_238":81,"_111":-5,"_120":73,"_122":20,"_133":294,"_124":20},{"_236":292,"_238":292,"_111":-5,"_120":73,"_122":20,"_133":293,"_124":20},{"_236":290,"_238":290,"_111":-5,"_120":73,"_122":20,"_133":291,"_124":20},{"_236":288,"_238":288,"_111":-5,"_120":73,"_122":20,"_133":289,"_124":20},{"_236":286,"_238":286,"_111":-5,"_120":73,"_122":20,"_133":287,"_124":20},{"_236":284,"_238":284,"_111":-5,"_120":73,"_122":20,"_133":285,"_124":20},{"_236":282,"_238":282,"_111":-5,"_120":73,"_122":20,"_133":283,"_124":20},{"_236":280,"_238":280,"_111":-5,"_120":73,"_122":20,"_133":281,"_124":20},{"_236":278,"_238":278,"_111":-5,"_120":73,"_122":20,"_133":279,"_124":20},{"_236":276,"_238":276,"_111":-5,"_120":73,"_122":20,"_133":277,"_124":20},{"_236":274,"_238":274,"_111":-5,"_120":73,"_122":20,"_133":275,"_124":20},{"_236":272,"_238":272,"_111":-5,"_120":73,"_122":20,"_133":273,"_124":20},{"_236":270,"_238":270,"_111":-5,"_120":73,"_122":20,"_133":271,"_124":20},{"_236":268,"_238":268,"_111":-5,"_120":73,"_122":20,"_133":269,"_124":20},{"_236":266,"_238":266,"_111":-5,"_120":73,"_122":20,"_133":267,"_124":20},{"_236":264,"_238":264,"_111":-5,"_120":73,"_122":20,"_133":265,"_124":20},{"_236":262,"_238":262,"_111":-5,"_120":73,"_122":20,"_133":263,"_124":20},{"_236":260,"_238":260,"_111":-5,"_120":73,"_122":20,"_133":261,"_124":20},{"_236":258,"_238":258,"_111":-5,"_120":73,"_122":20,"_133":259,"_124":20},{"_236":256,"_238":256,"_111":-5,"_120":73,"_122":20,"_133":257,"_124":20},{"_236":254,"_238":254,"_111":-5,"_120":73,"_122":20,"_133":255,"_124":20},{"_236":252,"_238":252,"_111":-5,"_120":73,"_122":20,"_133":253,"_124":20},{"_236":250,"_238":250,"_111":-5,"_120":73,"_122":20,"_133":251,"_124":20},{"_236":248,"_238":248,"_111":-5,"_120":73,"_122":20,"_133":249,"_124":20},{"_236":246,"_238":246,"_111":-5,"_120":73,"_122":20,"_133":247,"_124":20},{"_236":244,"_238":244,"_111":-5,"_120":73,"_122":20,"_133":245,"_124":20},{"_236":242,"_238":242,"_111":-5,"_120":73,"_122":20,"_133":243,"_124":20},{"_236":240,"_238":240,"_111":-5,"_120":73,"_122":20,"_133":241,"_124":20},{"_236":237,"_238":237,"_111":-5,"_120":73,"_122":20,"_133":239,"_124":20},"id","b200x1-throughput-fp8-ueeogrvolw","name",[],"b200x1-throughput-nvfp4-lzk6scakha",[],"b200x2-latency-bf16-srzmz11lsq",[],"b200x2-latency-fp8-kwktcc65kw",[],"b200x2-latency-nvfp4-akj4unhong",[],"gb200x1-throughput-bf16-b7mc0n5vsq",[],"gb200x1-throughput-fp8-eqspsgvc4g",[],"gb200x1-throughput-nvfp4-yqo6gpzgtw",[],"gb200x2-latency-bf16-ql0dncdzug",[],"gb200x2-latency-fp8-hh-qiitbsq",[],"gb200x2-latency-nvfp4-3sifj870lq",[],"gh200-144gbx1-throughput-bf16-hj8vo2ht-w",[],"gh200-144gbx1-throughput-fp8-xl4fpuq7yg",[],"gh200-144gbx2-latency-bf16-bacrp0khpq",[],"gh200-144gbx2-latency-fp8-9nydmnmc-g",[],"gh200-480gbx1-latency-fp8-5cegi-nk3a",[],"gh200-480gbx1-throughput-fp8-e010yvxmig",[],"h100-nvlx1-throughput-fp8-o6gx8o9qrw",[],"h100-nvlx2-latency-fp8-wllza3qfna",[],"h100-nvlx2-throughput-bf16-pfikm8whig",[],"h100-nvlx4-latency-bf16-2t19rh-jfa",[],"h100x1-throughput-fp8-zmf7sc5wtg",[],"h100x2-latency-fp8-sowqqe--5a",[],"h100x2-throughput-bf16-sfp5psfsoa",[],"h100x4-latency-bf16-c7ags8vtqa",[],"h200-nvlx1-throughput-bf16-2fchm5wb4g",[],"h200-nvlx1-throughput-fp8-is7q-dw1mq",[],"h200-nvlx2-latency-bf16-f-nnvkkyfa",[],[],"h200x1-throughput-bf16-jvvivxsong",[],"h200x1-throughput-fp8-tnrs6lwhqg",[],"h200x2-latency-fp8-qionglmjjw",[],"h200x4-latency-bf16-gwrdidufkg",[],"l40sx4-latency-bf16-fcnx7qsagw",[],"l40sx4-latency-fp8-csbsvltszw",[],"l40sx4-throughput-bf16-lb51ks7uxa",[],"l40sx4-throughput-fp8-grh3fk4vxa",[],"rtx6000-blackwell-svx1-throughput-fp8-naq8xshwuw",[],"rtx6000-blackwell-svx1-throughput-nvfp4-xavjkqnlyg",[],"rtx6000-blackwell-svx2-latency-fp8-dmfkw2yokg",[],"rtx6000-blackwell-svx2-latency-nvfp4-nxskuzhqig",[],"rtx6000-blackwell-svx2-throughput-bf16-dhgigcxxoq",[],"rtx6000-blackwell-svx2-throughput-nvfp4-pwfx1fcmsq",[],"rtx6000-blackwell-svx4-latency-bf16-ymougp5ifg",[],"fp8-rmctserfba-tool-calling",[],"nvfp4-biyprhc44a-tool-calling",[],"trtllmapi-pt-runtime-params-a100-sxm4-40gbx4-throughput-bf16-kiai9xwefa",[],"trtllmapi-pt-runtime-params-a100-sxm4-40gbx8-latency-bf16-3rxkym8ltq",[],"trtllmapi-pt-runtime-params-a100x2-throughput-bf16-nkaanvi0ia",[],"trtllmapi-pt-runtime-params-a100x4-latency-bf16-9asnlkhgrq",[],"trtllmapi-pt-runtime-params-a10gx8-latency-bf16-8r30w1bbyg",[],"trtllmapi-pt-runtime-params-a10gx8-throughput-bf16-ocrn2kfqua",[],"trtllmapi-pt-runtime-params-b200x1-throughput-bf16-soskusgelw",[],"trtllmapi-pt-runtime-params-b200x1-throughput-fp8-jc0m93mazw",[],"trtllmapi-pt-runtime-params-b200x1-throughput-nvfp4-eukrtsmgaa",[],"trtllmapi-pt-runtime-params-b200x2-latency-bf16-sww2a-etyg",[],"trtllmapi-pt-runtime-params-b200x2-latency-fp8-upwwqgxuqg",[],"trtllmapi-pt-runtime-params-b200x2-latency-nvfp4-7bv7ojw4qq",[],"trtllmapi-pt-runtime-params-gb200x1-throughput-bf16-awc6geunfg",[],"trtllmapi-pt-runtime-params-gb200x1-throughput-fp8-odcmmzkweq",[],"trtllmapi-pt-runtime-params-gb200x1-throughput-nvfp4-6jwkuwj7wa",[],"trtllmapi-pt-runtime-params-gb200x2-latency-bf16-uxsyjwgyiw",[],"trtllmapi-pt-runtime-params-gb200x2-latency-fp8-2w-q9o8b-q",[],"trtllmapi-pt-runtime-params-gb200x4-latency-nvfp4-8j099rkoxg",[],"trtllmapi-pt-runtime-params-gh200-144gbx1-throughput-bf16-ygblfcgflw",[],"trtllmapi-pt-runtime-params-gh200-144gbx1-throughput-fp8-64-gon9qcq",[],"trtllmapi-pt-runtime-params-gh200-144gbx2-latency-bf16-3mbbc-7ajg",[],"trtllmapi-pt-runtime-params-gh200-144gbx2-latency-fp8-7w6si5eb5w",[],"trtllmapi-pt-runtime-params-gh200-480gbx1-latency-fp8-xdmojjj16a",[],"trtllmapi-pt-runtime-params-gh200-480gbx1-throughput-fp8-95gzxium1a",[],"trtllmapi-pt-runtime-params-h100-nvlx1-throughput-fp8-ax-85sarrq",[],"trtllmapi-pt-runtime-params-h100-nvlx2-latency-fp8-plxydi5lcw",[],"trtllmapi-pt-runtime-params-h100-nvlx2-throughput-bf16-dtkwlmgxoa",[],"trtllmapi-pt-runtime-params-h100-nvlx4-latency-bf16-oibhi1q6oa",[],"trtllmapi-pt-runtime-params-h100x1-throughput-fp8-vfwiu03deg",[],"trtllmapi-pt-runtime-params-h100x2-latency-fp8-opxrciek2g",[],"trtllmapi-pt-runtime-params-h100x2-throughput-bf16-jsid2ao4yw",[],"trtllmapi-pt-runtime-params-h100x4-latency-bf16-7ju0k825oq",[],"trtllmapi-pt-runtime-params-h200-nvlx1-throughput-bf16-w3oymyqdmq",[],"trtllmapi-pt-runtime-params-h200-nvlx1-throughput-fp8-n0h4rfwxyq",[],"trtllmapi-pt-runtime-params-h200-nvlx2-latency-bf16-mv44kxbgxq",[],"trtllmapi-pt-runtime-params-h200-nvlx2-latency-fp8-mgp8axr5cg",[],"trtllmapi-pt-runtime-params-h200x1-throughput-bf16-r34k3h7cka",[],"trtllmapi-pt-runtime-params-h200x1-throughput-fp8-byat5hkiuw",[],"trtllmapi-pt-runtime-params-h200x2-latency-bf16-c1s1ehheoq",[],"trtllmapi-pt-runtime-params-h200x2-latency-fp8-nogiuyoqqg",[],"trtllmapi-pt-runtime-params-l40sx4-latency-bf16-qvqsvcxudq",[],"trtllmapi-pt-runtime-params-l40sx4-latency-fp8-igxoyg33gq",[],"trtllmapi-pt-runtime-params-l40sx4-throughput-bf16-qv0sdx9f-w",[],"trtllmapi-pt-runtime-params-l40sx4-throughput-fp8-v0418ijixq",[],"trtllmapi-pt-runtime-params-rtx6000-blackwell-svx1-throughput-fp8-yxpg7-lwrq",[],"trtllmapi-pt-runtime-params-rtx6000-blackwell-svx1-throughput-nvfp4-q-nc1wgrpg",[],"trtllmapi-pt-runtime-params-rtx6000-blackwell-svx2-latency-fp8-glfz-vixvw",[],"trtllmapi-pt-runtime-params-rtx6000-blackwell-svx2-latency-nvfp4-v2qlz-s4ua",[],"trtllmapi-pt-runtime-params-rtx6000-blackwell-svx4-throughput-bf16-wzczcd7neg",[],"trtllmapi-pt-runtime-params-rtx6000-blackwell-svx8-latency-bf16-t7fpezcdtw",[],"hf-checkpoint-bf16-jetart",[],"hf-checkpoint-bf16-jetart-toolparser-fix-v2",[],"fp8-rmctserfba-tool-calling-toolparser-fix-v2",[],"nvfp4-biyprhc44a-tool-calling-toolparser-fix-v2",[],"-","Latest (nvfp4-biyprhc44a-tool-calling-toolparser-fix-v2)","nim/nvidia/llama-3.3-nemotron-super-49b-v1.5","key","nim/nvidia/models/llama-3.3-nemotron-super-49b-v1.5","resourceType","models","resourceId","llama-3.3-nemotron-super-49b-v1.5","shortDescription","Llama-3.3-Nemotron-Super-49B-v1.5 is a significantly upgraded version of Llama-3.3-Nemotron-Super-49B-v1 and is a large language model (LLM) which is a derivative of Meta Llama-3.3-70B-Instruct","## **Llama-3.3-Nemotron-Super-49B-v1.5 Overview**\n\n## **Description:**\n\n**Llama-3.3-Nemotron-Super-49B-v1.5** is a significantly upgraded version of Llama-3.3-Nemotron-Super-49B-v1 and is a large language model (LLM) which is a derivative of Meta Llama-3.3-70B-Instruct (AKA the reference model). It is a reasoning model that is post trained for reasoning, human chat preferences, and agentic tasks, such as Retrieval-Augmented Generation (RAG) and tool calling. The model supports a context length of 128K tokens.\n\nLlama-3.3-Nemotron-Super-49B-v1.5 is a model which offers a great tradeoff between model accuracy and efficiency. Efficiency (throughput) directly translates to savings. Using a novel Neural Architecture Search (NAS) approach, we greatly reduce the model’s memory footprint, enabling larger workloads, as well as fitting the model on a single GPU at high workloads (H200). This NAS approach enables the selection of a desired point in the accuracy-efficiency tradeoff. For more information on the NAS approach, please refer to [this paper](https://arxiv.org/abs/2411.19146).\n\nThe model underwent a multi-phase post-training process to enhance both its reasoning and non-reasoning capabilities. This includes a supervised fine-tuning stage for Math, Code, Science, and Tool Calling. Additionally, the model went through multiple stages of Reinforcement Learning (RL) including Reward-aware Preference Optimization (RPO) for chat, Reinforcement Learning with Verifiable Rewards (RLVR) for reasoning, and iterative Direct Preference Optimization (DPO) for Tool Calling capability enhancements. The final checkpoint was achieved after merging several RL and DPO checkpoints.\n\nThis model is part of the Llama Nemotron Collection. You can find the other model(s) in this family here:\n\n* [Llama-3.1-Nemotron-Nano-4B-v1.1](https://huggingface.co/nvidia/Llama-3.1-Nemotron-Nano-4B-v1.1) \n* [Llama-3.1-Nemotron-Ultra-253B-v1](https://huggingface.co/nvidia/Llama-3_1-Nemotron-Ultra-253B-v1)\n\nThis model is ready for commercial use.\n\n## **License/Terms of Use**\n\n**GOVERNING TERMS:** The NIM container is governed by the [NVIDIA Software License Agreement](https://www.nvidia.com/en-us/agreements/enterprise-software/nvidia-software-license-agreement/) and the [Product-Specific Terms for NVIDIA AI Products](https://www.nvidia.com/en-us/agreements/enterprise-software/product-specific-terms-for-ai-products/); and the use of this model is governed by the [NVIDIA Open Model License](https://www.nvidia.com/en-us/agreements/enterprise-software/nvidia-open-model-license/).\n\n**Additional Information:** [Llama 3.3 Community License Agreement](https://www.llama.com/llama3_3/license/). Built with Llama.\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**You are responsible for ensuring that your use of NVIDIA provided models complies with all applicable laws.**\n\n## **Deployment Geography:**\n\nGlobal\n\n## **Use Case:**\n\nLlama-3.3-Nemotron-Super-49B-v1.5 is a general purpose reasoning and chat model intended to be used in English and coding languages. Other non-English languages (German, French, Italian, Portuguese, Hindi, Spanish, and Thai) are also supported. Developers designing AI Agent systems, chatbots, RAG systems, and other AI-powered applications. Also suitable for typical instruction-following tasks.\n\n## **Release Date:**\n\n* Hugging Face 7/25/2025 via [Llama-3\\_3-Nemotron-Super-49B-v1\\_5](https://huggingface.co/nvidia/Llama-3_3-Nemotron-Super-49B-v1_5) \n* Build.NVIDIA.com 7/25/2025 [Llama-3\\_3-Nemotron-Super-49B-v1\\_5](https://build.nvidia.com/nvidia/llama-3_3-nemotron-super-49b-v1_5)\n\n## **References**\n\n* [Llama-Nemotron: Efficient Reasoning Models (arXiv:2505.00949)](https://arxiv.org/abs/2505.00949) \n* [\\[2502.00203\\] Reward-aware Preference Optimization: A Unified Mathematical Framework for Model Alignment](https://arxiv.org/abs/2502.00203) \n* [\\[2411.19146\\]Puzzle: Distillation-Based NAS for Inference-Optimized LLMs](https://arxiv.org/abs/2411.19146)\n\n## **Model Architecture**\n\n**Architecture Type:** Dense decoder-only Transformer model\n\n**Network Architecture:** Llama 3.3 70B Instruct, customized through Neural Architecture Search (NAS)\n\nThe model is a derivative of Meta’s Llama-3.3-70B-Instruct, using Neural Architecture Search (NAS). The NAS algorithm results in non-standard and non-repetitive blocks. This includes the following:\n\nSkip attention: In some blocks, the attention is skipped entirely, or replaced with a single linear layer. Variable FFN: The expansion/compression ratio in the FFN layer is different between blocks.\n\nWe utilize a block-wise distillation of the reference model, where for each block we create multiple variants providing different tradeoffs of quality vs. computational complexity, discussed in more depth below. We then search over the blocks to create a model which meets the required throughput and memory (optimized for a single H100-80GB GPU) while minimizing the quality degradation. The model then undergoes knowledge distillation (KD), with a focus on English single and multi-turn chat use cases. The KD step included 40 billion tokens consisting of a mixture of 3 datasets \\- FineWeb, Buzz-V1.2, and Dolma.\n\n**Intended Use** \nLlama-3.3-Nemotron-Super-49B-v1.5 is a general purpose reasoning and chat model intended to be used in English and coding languages. Other non-English languages (German, French, Italian, Portuguese, Hindi, Spanish, and Thai) are also supported.\n\n## **Input**\n\n* **Input Type(s):** Text \n* **Input Format:** String \n* **Input Parameters:** One-Dimensional (1D) \n* **Other Properties Related to Input:** Context length up to 131,072 tokens\n\n## **Output**\n\n* **Output Type(s):** Text \n* **Output Format:** String \n* **Output Parameters:** One-Dimensional (1D) \n* **Other Properties Related to Output:** Context length up to 131,072 tokens\n\nOur AI models are designed and/or 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:** vLLM, TensorRT \n* **Supported Hardware Microarchitecture Compatibility:** \n * NVIDIA Ampere \n * NVIDIA Hopper \n* **Preferred 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:**\n\nLlama-3.3-Nemotron-Super-49B-v1.5-1.12.0
\nLlama-3.3-Nemotron-Super-49B-v1.5-1.13.1
\nLlama-3.3-Nemotron-Super-49B-v1.5-1.14.0
\nLlama-3.3-Nemotron-Super-49B-v1.5-1.15.0
\nLlama-3.3-Nemotron-Super-49B-v1.5-1.15.5
\n\n## **Quick Start and Usage Recommendations:**\n\n* By default (empty system prompt), the model will respond in reasoning ON mode. Setting /no\\_think in the system prompt will enable reasoning OFF mode. \n* We recommend setting temperature to 0.6 and Top P to 0.95 for Reasoning ON mode. \n* We recommend using greedy decoding for Reasoning OFF mode.\n\nYou can try this model out through the preview API, using this link: [Llama-3\\_3-Nemotron-Super-49B-v1\\_5](https://build.nvidia.com/nvidia/llama-3_3-nemotron-super-49b-v1_5).\n\n**Use It with vLLM**\n\npip install vllm==0.9.2\n\nAn example on how to serve with vLLM:\n\n```\n$ python3 -m vllm.entrypoints.openai.api_server \\\n --model \"nvidia/Llama-3_3-Nemotron-Super-49B-v1_5\" \\\n --trust-remote-code \\\n --seed=1 \\\n --host=\"0.0.0.0\" \\\n --port=5000 \\\n --served-model-name \"Llama-3_3-Nemotron-Super-49B-v1_5\" \\\n --tensor-parallel-size=8 \\\n --max-model-len=65536 \\\n --gpu-memory-utilization 0.95 \\\n --enforce-eager\n```\n\n**Running a vLLM Server with Tool-call Support**\n\nTo enable tool calling usage with this model, we provide a tool parser in the repository. Here is an example on how to use it:\n\n```\n$ git clone https://huggingface.co/nvidia/Llama-3_3-Nemotron-Super-49B-v1_5\n\n$ conda create -n vllm python=3.12 -y\n$ conda activate vllm\n$ pip install vllm==0.9.2\n\n$ python3 -m vllm.entrypoints.openai.api_server \\\n --model Llama-3_3-Nemotron-Super-49B-v1_5 \\\n --trust-remote-code \\\n --seed=1 \\\n --host=\"0.0.0.0\" \\\n --port=5000 \\\n --served-model-name \"Llama-3_3-Nemotron-Super-49B-v1_5\" \\\n --tensor-parallel-size=8 \\\n --max-model-len=65536 \\\n --gpu-memory-utilization 0.95 \\\n --enforce-eager \\\n --enable-auto-tool-choice \\\n --tool-parser-plugin \"Llama-3_3-Nemotron-Super-49B-v1_5/llama_nemotron_toolcall_parser_no_streaming.py\" \\\n --tool-call-parser \"llama_nemotron_json\"\n```\n\nAfter launching a vLLM server, you can call the server with tool-call support using a Python script like below.\n\n```py\nfrom openai import OpenAI\nclient = OpenAI(\n base_url=\"http://0.0.0.0:5000/v1\",\n api_key=\"dummy\",\n)\ncompletion = client.chat.completions.create(\n model=\"Llama-3_3-Nemotron-Super-49B-v1_5\",\n messages=[\n {\"role\": \"system\", \"content\": \"\"},\n {\"role\": \"user\", \"content\": \"My bill is $100. What will be the amount for 18% tip?\"}\n ],\n tools=[\n {\n \"type\": \"function\",\n \"function\": {\n \"name\": \"calculate_tip\",\n \"parameters\": {\n \"type\": \"object\",\n \"properties\": {\n \"bill_total\": {\n \"type\": \"integer\",\n \"description\": \"The total amount of the bill\"\n },\n \"tip_percentage\": {\n \"type\": \"integer\",\n \"description\": \"The percentage of tip to be applied\"\n }\n },\n \"required\": [\"bill_total\", \"tip_percentage\"]\n }\n }\n },\n {\n \"type\": \"function\",\n \"function\": {\n \"name\": \"convert_currency\",\n \"parameters\": {\n \"type\": \"object\",\n \"properties\": {\n \"amount\": {\n \"type\": \"integer\",\n \"description\": \"The amount to be converted\"\n },\n \"from_currency\": {\n \"type\": \"string\",\n \"description\": \"The currency code to convert from\"\n },\n \"to_currency\": {\n \"type\": \"string\",\n \"description\": \"The currency code to convert to\"\n }\n },\n \"required\": [\"from_currency\", \"amount\", \"to_currency\"]\n }\n }\n }\n ],\n temperature=0.6,\n top_p=0.95,\n max_tokens=32768,\n stream=False\n)\nprint(completion.choices[0].message.content)\n'''\n\nOkay, let's see. The user has a bill of $100 and wants to know the amount for an 18% tip. Hmm, I need to calculate the tip based on the bill total and the percentage. The tools provided include calculate_tip, which takes bill_total and tip_percentage as parameters. So the bill_total here is 100, and the tip_percentage is 18. I should call the calculate_tip function with these values. Wait, do I need to check if the parameters are integers? The bill is $100, which is an integer, and 18% is also an integer. So that fits the function's requirements. I don't need to convert any currency here because the user is asking about a tip in the same currency. So the correct tool to use is calculate_tip with those parameters.\n\n'''\nprint(completion.choices[0].message.tool_calls)\n'''\n[ChatCompletionMessageToolCall(id='chatcmpl-tool-e341c6954d2c48c2a0e9071c7bdefd8b', function=Function(arguments='{\"bill_total\": 100, \"tip_percentage\": 18}', name='calculate_tip'), type='function')]\n'''\n```\n\n## **Training, Testing, and Evaluation Datasets:**\n\n### **Training Dataset:**\n\nA large variety of training data was used for the knowledge distillation phase before the post-training pipeline, 3 of which included: FineWeb, Buzz-V1.2, and Dolma.\n\nThe data for the multi-stage post-training phases for improvements in Code, Math, and Reasoning is a compilation of SFT and RL data that supports improvements of math, code, general reasoning, and instruction following capabilities of the original Llama instruct model.\n\nPrompts have been sourced from either public and open corpus or synthetically generated. Responses were synthetically generated by a variety of models, with some prompts containing responses for both reasoning on and off modes, to train the model to distinguish between two modes.\n\nNVIDIA will be releasing the post-training dataset in the coming weeks.\n\n**Data Modality:** Text \n\n**Data Collection Method by dataset:** Hybrid: Automated, Human, Synthetic\n\n**Labeling Method by dataset:** Hybrid: Automated, Human, Synthetic\n\n**Properties (Quantity, Dataset Descriptions, Sensor(s)):** \n\nQuantity: The knowledge distillation phase used 40 billion tokens. The underlying Dolma dataset contains approximately 3 trillion tokens, and FineWeb contains over 15 trillion tokens.\n\nDataset Descriptions: The training data is a composite of large-scale web text, synthetically generated conversational data for coding and instruction-following, and a diverse corpus of text from academic, literary, and encyclopedic sources. The model supports a context length of 128,000 tokens.\n\n### \n\n### **Testing Dataset:**\n\n**Data Collection Method by dataset:** Hybrid: Human, Synthetic\n\n**Labeling Method by dataset:** Hybrid: Human, Automated\n\n**Properties (Quantity, Dataset Descriptions, Sensor(s)):**\n\nQuantity: The benchmarks contain thousands of test items, including:\n\n* MMLU: \\~15,900 multiple-choice questions. \n* GPQA: 448 difficult multiple-choice questions. \n* HumanEval: 164 programming problems. \n* GSM-8K: \\~8,500 grade-school math problems. \n* MATH: 12,500 competition math problems. \n* IF-Eval: Over 400 instruction-following prompts.\n\nDataset Descriptions: The evaluation suite is diverse, covering 57 academic and professional subjects (MMLU), expert-level reasoning (GPQA), Python code generation (HumanEval), mathematical problem-solving (GSM-8K, MATH), and the ability to follow precise instructions (IF-Eval).\n\n### **Evaluation Dataset:**\n\nWe used the datasets listed below to evaluate Llama-3.3-Nemotron-Super-49B-v1.5.\n\n**Data Collection Method by dataset:** Hybrid: Human. Synthetic\n\n**Labeling Method by dataset:** Hybrid: Human, Synthetic, Automatic\n\n**Evaluation Results** \nWe evaluate the model using temperature=0.6, top\\_p=0.95, and 64k sequence length. We run the benchmarks up to 16 times and average the scores to be more accurate.\n\n### MATH500\n\n| Reasoning Mode | pass@1 (avg. over 4 runs) |\n| ----- | ----- |\n| Reasoning On | 97.4 |\n\n### AIME 2024\n\n| Reasoning Mode | pass@1 (avg. over 16 runs) |\n| ----- | ----- |\n| Reasoning On | 87.5 |\n\n### AIME 2025\n\n| Reasoning Mode | pass@1 (avg. over 16 runs) |\n| ----- | ----- |\n| Reasoning On | 82.71 |\n\n### GPQA\n\n| Reasoning Mode | pass@1 (avg. over 4 runs) |\n| ----- | ----- |\n| Reasoning On | 71.97 |\n\n### LiveCodeBench 24.10-25.02\n\n| Reasoning Mode | pass@1 (avg. over 4 runs) |\n| ----- | ----- |\n| Reasoning On | 73.58 |\n\n### BFCL v3\n\n| Reasoning Mode | pass@1 (avg. over 2 runs) |\n| ----- | ----- |\n| Reasoning On | 71.75 |\n\n### IFEval\n\n| Reasoning Mode | Strict:Instruction |\n| ----- | ----- |\n| Reasoning On | 88.61 |\n\n### ArenaHard\n\n| Reasoning Mode | pass@1 (avg. over 1 runs) |\n| ----- | ----- |\n| Reasoning On | 92.0 |\n\n### Humanity's Last Exam (Text-Only Subset)\n\n| Reasoning Mode | pass@1 (avg. over 1 runs) |\n| ----- | ----- |\n| Reasoning On | 7.64 |\n\n### MMLU Pro (CoT)\n\n| Reasoning Mode | pass@1 (avg. over 1 runs) |\n| ----- | ----- |\n| Reasoning On | 79.53 |\n\nAll evaluations were done using the [NeMo-Skills](https://github.com/NVIDIA/NeMo-Skills) repository.\n\n## **Inference:**\n\n**Acceleration Engine:**\n\n* vLLM, TensorRT\n\n**Test Hardware:**\n\n* 2x NVIDIA H100-80GB \n* 2x NVIDIA A100-80GB GPUs\n* B200 SXM \n* H200 SXM \n* H100 SXM \n* A100 SXM 80GB \n* A100 SXM 40GB \n* L40S PCIe \n* A10G \n* H100 NVL \n* H200 NVL \n* GH200 96GB \n* GB200 \n* RTX 5090 \n* RTX 4090 \n* RTX 6000 Ada \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\nFor more detailed information on ethical considerations for this model, please see the Model Card++ [Bias](http://./bias.md), [Explainability](http://./explainability.md), [Safety & Security](http://./safety.md), and [Privacy](http://./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\nYou are responsible for ensuring that your use of NVIDIA provided models complies with all applicable laws.\n\n## **Citation**\n\n```py\n@misc{bercovich2025llamanemotronefficientreasoningmodels,\n title={Llama-Nemotron: Efficient Reasoning Models}, \n author={Akhiad Bercovich and Itay Levy and Izik Golan and Mohammad Dabbah and Ran El-Yaniv and Omri Puny and Ido Galil and Zach Moshe and Tomer Ronen and Najeeb Nabwani and Ido Shahaf and Oren Tropp and Ehud Karpas and Ran Zilberstein and Jiaqi Zeng and Soumye Singhal and Alexander Bukharin and Yian Zhang and Tugrul Konuk and Gerald Shen and Ameya Sunil Mahabaleshwarkar and Bilal Kartal and Yoshi Suhara and Olivier Delalleau and Zijia Chen and Zhilin Wang and David Mosallanezhad and Adi Renduchintala and Haifeng Qian and Dima Rekesh and Fei Jia and Somshubra Majumdar and Vahid Noroozi and Wasi Uddin Ahmad and Sean Narenthiran and Aleksander Ficek and Mehrzad Samadi and Jocelyn Huang and Siddhartha Jain and Igor Gitman and Ivan Moshkov and Wei Du and Shubham Toshniwal and George Armstrong and Branislav Kisacanin and Matvei Novikov and Daria Gitman and Evelina Bakhturina and Jane Polak Scowcroft and John Kamalu and Dan Su and Kezhi Kong and Markus Kliegl and Rabeeh Karimi and Ying Lin and Sanjeev Satheesh and Jupinder Parmar and Pritam Gundecha and Brandon Norick and Joseph Jennings and Shrimai Prabhumoye and Syeda Nahida Akter and Mostofa Patwary and Abhinav Khattar and Deepak Narayanan and Roger Waleffe and Jimmy Zhang and Bor-Yiing Su and Guyue Huang and Terry Kong and Parth Chadha and Sahil Jain and Christine Harvey and Elad Segal and Jining Huang and Sergey Kashirsky and Robert McQueen and Izzy Putterman and George Lam and Arun Venkatesan and Sherry Wu and Vinh Nguyen and Manoj Kilaru and Andrew Wang and Anna Warno and Abhilash Somasamudramath and Sandip Bhaskar and Maka Dong and Nave Assaf and Shahar Mor and Omer Ullman Argov and Scot Junkin and Oleksandr Romanenko and Pedro Larroy and Monika Katariya and Marco Rovinelli and Viji Balas and Nicholas Edelman and Anahita Bhiwandiwalla and Muthu Subramaniam and Smita Ithape and Karthik Ramamoorthy and Yuting Wu and Suguna Varshini Velury and Omri Almog and Joyjit Daw and Denys Fridman and Erick Galinkin and Michael Evans and Katherine Luna and Leon Derczynski and Nikki Pope and Eileen Long and Seth Schneider and Guillermo Siman and Tomasz Grzegorzek and Pablo Ribalta and Monika Katariya and Joey Conway and Trisha Saar and Ann Guan and Krzysztof Pawelec and Shyamala Prayaga and Oleksii Kuchaiev and Boris Ginsburg and Oluwatobi Olabiyi and Kari Briski and Jonathan Cohen and Bryan Catanzaro and Jonah Alben and Yonatan Geifman and Eric Chung and Chris Alexiuk},\n year={2025},\n eprint={2505.00949},\n archivePrefix={arXiv},\n primaryClass={cs.CL},\n url={https://arxiv.org/abs/2505.00949}, \n}\n```","displayName","Llama-3.3-Nemotron-Super-49B-v1.5",["D",1785885023354],"publisher","NVIDIA","latestVersion","versions","collectionCounts","imageUrl","/api/image/nim/nvidia/models/llama-3.3-nemotron-super-49b-v1.5","rawImageUrl","https://catalog.ngc.nvidia.com/_next/image?url=https%3A%2F%2Fassets.ngc.nvidia.com%2Fproducts%2Fapi-catalog%2Fimages%2Fllama-3_3-nemotron-49b-instruct.jpg&w=640&q=90","canGuestDownload","latestVersionSizeInBytes",31090578372,"licenseTerms",{"_512":513,"_514":515,"_516":517,"_518":20,"_519":520},"isMultinodeEnabled","isMultiArchitecture","bias","### **Bias**\n\n|Field:|Response:|\n|:---:|:---:|\n|Participation considerations from adversely impacted groups (protected classes) in model design and testing:|None|\n|Measures taken to mitigate against unwanted bias:|None|","explainability","### **Explainability**\n\n| Field: | Response: |\n| :---- | :---- |\n| Intended Application(s) & Domain(s): | Text generation, reasoning, summarization, and question answering. |\n| Model Type: | Text-to-text transformer |\n| Intended Users: | This model is intended for developers, researchers, and customers building/utilizing LLMs, while balancing accuracy and efficiency. |\n| Output: | Text String(s) |\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| Technical Limitations: | The model was trained on data that contains toxic language, unsafe content, and societal biases originally crawled from the internet. Therefore, the model may amplify those biases and return toxic responses especially when prompted with toxic prompts. The model may generate answers that may be inaccurate, omit key information, or include irrelevant or redundant text producing socially unacceptable or undesirable text, even if the prompt itself does not include anything explicitly offensive. The model demonstrates weakness to alignment-breaking attacks. Users are advised to deploy language model guardrails alongside this model to prevent potentially harmful outputs. The Model may generate answers that are inaccurate, omit key information, or include irrelevant or redundant text. |\n| Verified to have met prescribed 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 susceptible to prompt injection and jailbreaking in various forms as a result of its instruction tuning. 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 was trained on data that contains toxic language and societal biases originally crawled from the internet. Therefore, the model may amplify those biases and return toxic responses especially when prompted with toxic prompts. The model may generate answers that may be inaccurate, omit key information, or include irrelevant or redundant text producing socially unacceptable or undesirable text, even if the prompt itself does not include anything explicitly offensive. Model output should be appropriately escaped before viewing or other processing.|\n| End User License Agreement: | Your use of this model is governed by the [NVIDIA Software License Agreement](https://www.nvidia.com/en-us/agreements/enterprise-software/nvidia-software-license-agreement/) and the [Product-Specific Terms for NVIDIA AI Products](https://www.nvidia.com/en-us/agreements/enterprise-software/product-specific-terms-for-ai-products/); and the use of this model is governed by the [NVIDIA Open Model License](https://www.nvidia.com/en-us/agreements/enterprise-software/nvidia-open-model-license/). Additional Information: [Llama 3.3 Community License Agreement](https://www.llama.com/llama3_3/license/). Built with Llama. |","privacy","### Privacy\n\n|Field:|Response:|\n|:---:|:---:|\n|Generatable or Reverse engineerable personally-identifiable information?|None|\n|Was consent obtained for any personal data used?|None Known|\n|Personal data used to create this model?|None Known|\n|How often is dataset reviewed?|Before Release|\n|Is there provenance for all datasets used in training?|Yes|\n|Does data labeling (annotation, metadata) comply with privacy laws?|Yes|\n|Applicable NVIDIA Privacy Policy|https://www.nvidia.com/en-us/about-nvidia/privacy-policy/|","safetyAndSecurity","### Safety & Security\n\n|Field:|Response:|\n|:---:|:---:|\n|Model Application(s):|Chat, Instruction Following, Chatbot Development, Code Generation, Reasoning|\n|Describe life critical application (if present):|None Known (please see referenced Known Risks in the Explainability subcard).|\n|Use Case Restrictions:|Your use of this model is governed by the [NVIDIA Software License Agreement](https://www.nvidia.com/en-us/agreements/enterprise-software/nvidia-software-license-agreement/) and the [Product-Specific Terms for NVIDIA AI Products](https://www.nvidia.com/en-us/agreements/enterprise-software/product-specific-terms-for-ai-products/); and the use of this model is governed by the [NVIDIA Open Model License](https://www.nvidia.com/en-us/agreements/enterprise-software/nvidia-open-model-license/). Additional Information: [Llama 3.3 Community License Agreement](https://www.llama.com/llama3_3/license/). Built with Llama. |","featureLabels",[503,504],"displayLabels",[487,488,489],"policyLabels",[],"systemLabels",[],"products",[],"hasNimMetadata",{"_449":501,"_491":501,"_493":502,"_495":496},{"_449":499,"_491":499,"_493":500,"_495":496},{"_449":490,"_491":492,"_493":494,"_495":496},"NVIDIA AI","value","soln_nvidia_ai","filter",{"_497":498,"_491":492},"category","label","field","technology","NSPECT-8J48-AV9D",{"_497":496,"_491":499},"False",{"_497":496,"_491":501},{"_449":510,"_491":510,"_493":511,"_495":508},{"_449":505,"_491":506,"_493":507,"_495":508},"NVIDIA NIM","nimmcro_nvidia_nim",{"_497":509,"_491":506},"feature","nvidia_nim","NVIDIA AI Enterprise Supported",{"_497":496,"_491":510},"governingTerms","

Governing Terms: Use of this model is governed by the NVIDIA Open Model License.

The underlying model is also licensed under the Llama 3.3 Community License Agreement. Built with Llama.

","licenseId","NVIDIA proprietary","licenseVersion","v1.0","needsAcceptance","configuredLicenseTerms",{"_521":522},"download",{"_523":524,"_525":526},"tags",[527],"generatedHtml","

Governing Terms: Use of this model is governed by the NVIDIA Open Model License.

The underlying model is also licensed under the Llama 3.3 Community License Agreement. Built with Llama.

",{"_440":528,"_495":529,"_530":531,"_496":531},"see-governing-terms","terms","tag","See Governing Terms","appConfig",{"_1513":1514,"_566":1515},"ngcApps",[1466],"csaApiBaseUrl","https://www.nvidia.com/ai-agents/csa-agent/dcsm","datadogRum",{"_1447":1448,"_1449":1450,"_1451":1452,"_1453":1454,"_1455":544,"_80":573,"_1456":1457,"_1458":44,"_1459":73,"_1460":73,"_1461":1462,"_1463":73,"_1464":1465},"amplitude",{"_1444":1445,"_1086":-7,"_1446":20},"disableJs","environment","prod","featureFlagsBlob","JTdCJTIydGVzdEZsYWclMjIlM0FmYWxzZSUyQyUyMnNob3dTZWFyY2hMYWJlbFNldEZpbHRlcnMlMjIlM0FmYWxzZSUyQyUyMmVuYWJsZUNzYVdpZGdldCUyMiUzQWZhbHNlJTJDJTIyZW5hYmxlUHJvZHVjdEdyb3VwaW5nQnlTdWJzY3JpcHRpb24lMjIlM0FmYWxzZSUyQyUyMmNvbnRhaW5lclNjYW5uaW5nUmVkZXNpZ25NMiUyMiUzQWZhbHNlJTdE","featureFlagsGuarded","isDevLikeEnv","isPlaywright","isMobileDevice","categorizedOrgs",{"_1438":1439,"_1440":1441,"_1442":1443},"currentOrgContext","currentContextRoles",[],"enterpriseOrgs",[],"manageableExternalUserGroups",[],"ngcApiOrigin","https://api.ngc.nvidia.com","ngcHelmOrigin","https://helm.ngc.nvidia.com","origin","https://catalog.ngc.nvidia.com","user","usesMocks","theme","year",2026,"now",1787766360570,"1.12.0","defaultSearchSuggestions",{"_1326":653,"_1327":1328,"_1329":1330,"_1331":1332},"announcements",[],"consentState",{"_611":1316,"_1317":73,"_1318":20,"_1319":20,"_1320":1321},"consentConfig",{"_593":594,"_595":596,"_597":598,"_599":600,"_601":602,"_603":604,"_605":606,"_607":608,"_609":610,"_611":612},"showContactAdmin",1787766360571,[],[],[587,588,589,590,591,592],{"_33":48,"_35":56,"_37":64,"_39":-5,"_41":-5,"_43":44},{"_33":48,"_35":54,"_37":64,"_39":-5,"_41":-5,"_43":44},{"_33":48,"_35":53,"_37":64,"_39":-5,"_41":-5,"_43":44},{"_33":48,"_35":49,"_37":64,"_39":-5,"_41":-5,"_43":44},{"_33":45,"_35":46,"_37":64,"_39":-5,"_41":-5,"_43":44},{"_33":34,"_35":36,"_37":38,"_39":40,"_41":47,"_43":44},"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",{"_1308":1309,"_1310":1311,"_1312":1313,"_1314":1315},"bannerAcceptLabel","Accept All","bannerRejectLabel","Reject Optional","bannerSettingsLabel","Manage Settings","preferenceSaveLabel","Save Choices","preferenceRejectLabel","Decline All","alwaysActiveText","Always Active","groups",[613,614,615,616,617],{"_618":997,"_238":998,"_109":999,"_622":73,"_623":1000,"_625":1001,"_627":1002},{"_618":900,"_238":901,"_109":902,"_622":20,"_623":42,"_625":903,"_627":904},{"_618":808,"_238":809,"_109":810,"_622":20,"_623":47,"_625":811,"_627":812},{"_618":629,"_238":630,"_109":631,"_622":20,"_623":119,"_625":632,"_627":633},{"_618":619,"_238":620,"_109":621,"_622":20,"_623":624,"_625":626,"_627":628},"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.",[750,751,752,753,754,755,756,757,758,759,760,761,762,763,764,765,766,767,768,769,770,771],[634,635,636,637,638,639,640,641,642,643,644,645,646,647,648,649,650],{"_651":745,"_449":745,"_109":653,"_625":746},{"_651":741,"_449":741,"_109":653,"_625":742},{"_651":736,"_449":736,"_109":653,"_625":737},{"_651":732,"_449":732,"_109":653,"_625":733},{"_651":727,"_449":727,"_109":653,"_625":728},{"_651":722,"_449":722,"_109":653,"_625":723},{"_651":714,"_449":714,"_109":653,"_625":715},{"_651":706,"_449":706,"_109":653,"_625":707},{"_651":701,"_449":701,"_109":653,"_625":702},{"_651":696,"_449":696,"_109":653,"_625":697},{"_651":691,"_449":691,"_109":653,"_625":692},{"_651":686,"_449":686,"_109":653,"_625":687},{"_651":681,"_449":681,"_109":653,"_625":682},{"_651":674,"_449":674,"_109":653,"_625":675},{"_651":669,"_449":669,"_109":653,"_625":670},{"_651":662,"_449":662,"_109":653,"_625":663},{"_651":652,"_449":652,"_109":653,"_625":654},"hostName","amazon-adsystem.com","",[655,656],{"_238":661,"_658":652,"_659":20,"_109":660},{"_238":657,"_658":652,"_659":20,"_109":660},"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",[664,665],{"_238":668,"_658":662,"_659":20,"_109":667},{"_238":666,"_658":662,"_659":20,"_109":667},"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",[671],{"_238":672,"_658":669,"_659":20,"_109":673},"MUID","This domain is owned by Mircosoft - it is the site for the search engine Bing.","assets.adobedtm.com",[676,677],{"_238":680,"_658":674,"_659":20,"_109":679},{"_238":678,"_658":674,"_659":20,"_109":679},"kndctr_F207D74D549850760A4C98C6_AdobeOrg_*","If you know anything about this host and its uses please add your contributions here. ","kndctr_F207D74D549850760A4C98C6_AdobeOrg_consent","tiktok.com",[683],{"_238":684,"_658":681,"_659":20,"_109":685},"_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",[688],{"_238":689,"_658":686,"_659":20,"_109":690},"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",[693],{"_238":694,"_658":691,"_659":20,"_109":695},"sp","Testing","google.com",[698],{"_238":699,"_658":696,"_659":20,"_109":700},"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",[703],{"_238":704,"_658":701,"_659":20,"_109":705},"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",[708,709],{"_238":712,"_658":706,"_659":20,"_109":713},{"_238":710,"_658":706,"_659":20,"_109":711},"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",[716,717],{"_238":720,"_658":714,"_659":20,"_109":721},{"_238":718,"_658":714,"_659":20,"_109":719},"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",[724],{"_238":725,"_658":722,"_659":20,"_109":726},"demdex","This cookie helps Adobe Audience Manger perform basic functions such as visitor identification, ID synchronization, segmentation, modeling, reporting, etc.","im-apps.net",[729],{"_238":730,"_658":727,"_659":20,"_109":731},"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",[734],{"_238":735,"_658":732,"_659":20,"_109":695},"R","6sc.co",[738],{"_238":739,"_658":736,"_659":20,"_109":740},"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",[743],{"_238":744,"_658":741,"_659":20,"_109":695},"_hly_gvid","t.contentsquare.net",[747],{"_238":748,"_658":745,"_659":73,"_109":749},"_cs_*","This is a session cookie. This is a pattern type cookie where a random 13-digit number is appended to the prefix _cs_.",{"_238":806,"_658":773,"_659":20,"_109":807},{"_238":804,"_658":773,"_659":20,"_109":805},{"_238":802,"_658":773,"_659":20,"_109":803},{"_238":684,"_658":773,"_659":20,"_109":801},{"_238":778,"_658":773,"_659":73,"_109":780},{"_238":800,"_658":773,"_659":20,"_109":653},{"_238":799,"_658":776,"_659":20,"_109":653},{"_238":797,"_658":773,"_659":20,"_109":798},{"_238":795,"_658":773,"_659":20,"_109":796},{"_238":794,"_658":773,"_659":20,"_109":653},{"_238":792,"_658":773,"_659":20,"_109":793},{"_238":778,"_658":791,"_659":73,"_109":780},{"_238":789,"_658":776,"_659":20,"_109":790},{"_238":789,"_658":779,"_659":20,"_109":790},{"_238":748,"_658":773,"_659":20,"_109":749},{"_238":787,"_658":773,"_659":20,"_109":788},{"_238":694,"_658":785,"_659":20,"_109":786},{"_238":783,"_658":773,"_659":20,"_109":784},{"_238":781,"_658":773,"_659":20,"_109":782},{"_238":778,"_658":779,"_659":73,"_109":780},{"_238":775,"_658":776,"_659":20,"_109":777},{"_238":772,"_658":773,"_659":73,"_109":774},"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.",[842,843,844,845,846,847,848,849,850,851,852,853,854,855,856,857,858,859,860,861,862,863,864,865,866],[813,814,815,816,817],{"_651":727,"_449":727,"_109":653,"_625":837},{"_651":701,"_449":701,"_109":653,"_625":832},{"_651":827,"_449":827,"_109":653,"_625":828},{"_651":822,"_449":822,"_109":653,"_625":823},{"_651":818,"_449":818,"_109":653,"_625":819},"nvidia.partners",[820],{"_238":821,"_658":818,"_659":20,"_109":695},"edge","app.hushly.com",[824],{"_238":825,"_658":822,"_659":20,"_109":826},"*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",[829],{"_238":830,"_658":827,"_659":20,"_109":831},"_octo","This domain is owned by Github. The main business activity is: Source code repository and collaboration platform",[833,834],{"_238":836,"_658":701,"_659":20,"_109":705},{"_238":835,"_658":701,"_659":20,"_109":705},"lidc","li_gc",[838,839],{"_238":841,"_658":727,"_659":20,"_109":731},{"_238":840,"_658":727,"_659":20,"_109":731},"imid_*","imid",{"_238":867,"_658":776,"_659":20,"_109":869},{"_238":898,"_658":773,"_659":20,"_109":899},{"_238":867,"_658":897,"_659":20,"_109":869},{"_238":895,"_658":773,"_659":73,"_109":896},{"_238":893,"_658":773,"_659":20,"_109":894},{"_238":892,"_658":875,"_659":73,"_109":653},{"_238":867,"_658":871,"_659":20,"_109":869},{"_238":890,"_658":773,"_659":20,"_109":891},{"_238":867,"_658":779,"_659":20,"_109":869},{"_238":878,"_658":776,"_659":20,"_109":653},{"_238":867,"_658":889,"_659":20,"_109":869},{"_238":887,"_658":773,"_659":20,"_109":888},{"_238":884,"_658":885,"_659":20,"_109":886},{"_238":882,"_658":883,"_659":20,"_109":653},{"_238":867,"_658":879,"_659":20,"_109":869},{"_238":881,"_658":773,"_659":20,"_109":653},{"_238":867,"_658":880,"_659":20,"_109":869},{"_238":878,"_658":879,"_659":20,"_109":653},{"_238":877,"_658":875,"_659":73,"_109":653},{"_238":867,"_658":876,"_659":20,"_109":869},{"_238":867,"_658":875,"_659":20,"_109":869},{"_238":867,"_658":874,"_659":20,"_109":869},{"_238":873,"_658":773,"_659":73,"_109":869},{"_238":870,"_658":871,"_659":73,"_109":872},{"_238":867,"_658":868,"_659":20,"_109":869},"_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.",[925,926,927,928,929,930,931,932,933,934,935,936,937,938,939,940,941,942,943,944,945,946,947,948,949,950,951,952,953],[905,906,907,908],{"_651":920,"_449":920,"_109":653,"_625":921},{"_651":915,"_449":915,"_109":653,"_625":916},{"_651":714,"_449":714,"_109":653,"_625":912},{"_651":827,"_449":827,"_109":653,"_625":909},[910],{"_238":911,"_658":827,"_659":73,"_109":831},"_gh_sess",[913],{"_238":914,"_658":714,"_659":73,"_109":721},"YSC","bizible.com",[917],{"_238":918,"_658":915,"_659":20,"_109":919},"_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",[922],{"_238":923,"_658":920,"_659":20,"_109":924},"tuid","This domain is owned by Ookla which provides internet speed testing services through its Speedtest application.",{"_238":995,"_658":773,"_659":20,"_109":996},{"_238":980,"_658":994,"_659":73,"_109":981},{"_238":963,"_658":776,"_659":20,"_109":888},{"_238":963,"_658":876,"_659":20,"_109":888},{"_238":992,"_658":773,"_659":20,"_109":993},{"_238":980,"_658":779,"_659":73,"_109":981},{"_238":990,"_658":773,"_659":20,"_109":991},{"_238":980,"_658":776,"_659":73,"_109":981},{"_238":988,"_658":875,"_659":20,"_109":989},{"_238":980,"_658":897,"_659":73,"_109":981},{"_238":960,"_658":776,"_659":20,"_109":962},{"_238":986,"_658":773,"_659":73,"_109":987},{"_238":984,"_658":773,"_659":20,"_109":985},{"_238":982,"_658":773,"_659":73,"_109":983},{"_238":980,"_658":871,"_659":73,"_109":981},{"_238":978,"_658":773,"_659":20,"_109":979},{"_238":976,"_658":773,"_659":20,"_109":977},{"_238":974,"_658":773,"_659":73,"_109":975},{"_238":972,"_658":880,"_659":20,"_109":973},{"_238":960,"_658":871,"_659":20,"_109":962},{"_238":970,"_658":773,"_659":20,"_109":971},{"_238":968,"_658":773,"_659":73,"_109":969},{"_238":966,"_658":773,"_659":20,"_109":967},{"_238":964,"_658":773,"_659":20,"_109":965},{"_238":963,"_658":773,"_659":20,"_109":888},{"_238":960,"_658":961,"_659":20,"_109":962},{"_238":958,"_658":773,"_659":20,"_109":959},{"_238":956,"_658":773,"_659":73,"_109":957},{"_238":954,"_658":773,"_659":20,"_109":955},"_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.",1,[1143,1144,1145,1146,1147,1148,1149,1150,1151,1152,1153,1154,1155,1156,1157,1158,1159,1160,1161,1162,1163,1164,1165,1166,1167,1168,1169,1170,1171,1172,1173,1174,1175,1176,1177,1178,1179,1180,1181,1182,1183,1184,1185,1186,1187,1188,1189,1190,1191,1192,1193,1194,1195,1196,1197,1198,1199,1200,1201,1202,1203,1204,1205,1206,1207,1208,1209,1210,1211,1212,1213,1214,1215],[1003,1004,1005,1006,1007,1008,1009,1010,1011,1012,1013,1014,1015,1016,1017,1018,1019,1020,1021,1022,1023,1024,1025,1026,1027],{"_651":1140,"_449":1140,"_109":653,"_625":1141},{"_651":1134,"_449":1134,"_109":653,"_625":1135},{"_651":1131,"_449":1131,"_109":653,"_625":1132},{"_651":1119,"_449":1119,"_109":653,"_625":1120},{"_651":1116,"_449":1116,"_109":653,"_625":1117},{"_651":1113,"_449":1113,"_109":653,"_625":1114},{"_651":1108,"_449":1108,"_109":653,"_625":1109},{"_651":714,"_449":714,"_109":653,"_625":1105},{"_651":1102,"_449":1102,"_109":653,"_625":1103},{"_651":706,"_449":706,"_109":653,"_625":1100},{"_651":701,"_449":701,"_109":653,"_625":1098},{"_651":1093,"_449":1093,"_109":653,"_625":1094},{"_651":1088,"_449":1088,"_109":653,"_625":1089},{"_651":1083,"_449":1083,"_109":653,"_625":1084},{"_651":1080,"_449":1080,"_109":653,"_625":1081},{"_651":827,"_449":827,"_109":653,"_625":1077},{"_651":686,"_449":686,"_109":653,"_625":1074},{"_651":822,"_449":822,"_109":653,"_625":1067},{"_651":818,"_449":818,"_109":653,"_625":1061},{"_651":1055,"_449":1055,"_109":653,"_625":1056},{"_651":1050,"_449":1050,"_109":653,"_625":1051},{"_651":1045,"_449":1045,"_109":653,"_625":1046},{"_651":1041,"_449":1041,"_109":653,"_625":1042},{"_651":1036,"_449":1036,"_109":653,"_625":1037},{"_651":1028,"_449":1028,"_109":653,"_625":1029},"polyfill.io",[1030,1031],{"_238":1034,"_658":1028,"_659":20,"_109":1035},{"_238":1032,"_658":1028,"_659":20,"_109":1033},"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",[1038],{"_238":1039,"_658":1036,"_659":20,"_109":1040},"_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",[1043],{"_238":1044,"_658":1041,"_659":20,"_109":695},"__cflb","instagram.com",[1047],{"_238":1048,"_658":1045,"_659":20,"_109":1049},"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",[1052],{"_238":1053,"_658":1050,"_659":20,"_109":1054},"__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",[1057,1058],{"_238":1060,"_658":1055,"_659":20,"_109":695},{"_238":1059,"_658":1055,"_659":20,"_109":695},"AWSALBTG","AWSALBTGCORS",[1062,1063],{"_238":1065,"_658":818,"_659":20,"_109":1066},{"_238":1064,"_658":818,"_659":20,"_109":695},"ak_bmsc","bm_mi","This domain is likely associated with NVIDIA's partner program, which could involve marketing, sales, and support services for partners.",[1068,1069,1070],{"_238":1073,"_658":822,"_659":20,"_109":695},{"_238":1072,"_658":822,"_659":73,"_109":695},{"_238":1071,"_658":822,"_659":20,"_109":695},"AWSALBCORS","JSESSIONID","AWSALB",[1075],{"_238":1053,"_658":686,"_659":20,"_109":1076},"This is a CloudFoundry cookie",[1078],{"_238":1079,"_658":827,"_659":20,"_109":831},"logged_in","hcaptcha.com",[1082],{"_238":1053,"_658":1080,"_659":20,"_109":1076},"tracking.intentsify.io",[1085],{"_238":1086,"_658":1083,"_659":20,"_109":1087},"userId","Used to track usage of the embedded video player.","static.reo.dev",[1090],{"_238":1091,"_658":1088,"_659":20,"_109":1092},"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",[1095],{"_238":1096,"_658":1093,"_659":20,"_109":1097},"dd_cookie_test_","dd_cookie_test",[1099],{"_238":1053,"_658":701,"_659":20,"_109":1076},[1101],{"_238":1053,"_658":706,"_659":20,"_109":1076},"supabase.co",[1104],{"_238":1053,"_658":1102,"_659":20,"_109":1076},[1106],{"_238":1107,"_658":714,"_659":20,"_109":721},"VISITOR_PRIVACY_METADATA","dosa-prd.azurewebsites.net",[1110],{"_238":1111,"_658":1108,"_659":20,"_109":1112},"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",[1115],{"_238":1053,"_658":1113,"_659":20,"_109":1076},"cdn.readme.io",[1118],{"_238":1053,"_658":1116,"_659":20,"_109":1076},"queue-it.net",[1121,1122,1123,1124,1125],{"_238":1130,"_658":1119,"_659":20,"_109":653},{"_238":1129,"_658":1119,"_659":20,"_109":653},{"_238":1128,"_658":1119,"_659":20,"_109":653},{"_238":1127,"_658":1119,"_659":20,"_109":653},{"_238":1126,"_658":1119,"_659":20,"_109":653},"Queue-it-{queueId}","Queue-it-{dmonvidia }________{eventId}","Queue-it-token","queueitsoftblock_{queueId}","Queue-it-{nvidia}________{eventId}","x.com",[1133],{"_238":1053,"_658":1131,"_659":20,"_109":1054},"api.hcaptcha.com",[1136,1137],{"_238":1139,"_658":1134,"_659":20,"_109":695},{"_238":1044,"_658":1134,"_659":20,"_109":1138},"This domain is associated with hCaptcha, a service that provides CAPTCHA and reCAPTCHA alternatives for user verification and bot protection.","hmt_id","vimeo.com",[1142],{"_238":1053,"_658":1140,"_659":20,"_109":1076},{"_238":1276,"_658":773,"_659":20,"_109":1307},{"_238":1305,"_658":868,"_659":73,"_109":1306},{"_238":1298,"_658":1267,"_659":73,"_109":1299},{"_238":1304,"_658":961,"_659":73,"_109":653},{"_238":1302,"_658":773,"_659":20,"_109":1303},{"_238":1060,"_658":1288,"_659":20,"_109":1227},{"_238":1235,"_658":1297,"_659":73,"_109":1237},{"_238":1059,"_658":1226,"_659":20,"_109":1227},{"_238":1238,"_658":1297,"_659":20,"_109":1239},{"_238":1300,"_658":773,"_659":73,"_109":1301},{"_238":1298,"_658":1274,"_659":73,"_109":1299},{"_238":1232,"_658":1297,"_659":20,"_109":1234},{"_238":1295,"_658":773,"_659":20,"_109":1296},{"_238":1059,"_658":1218,"_659":20,"_109":1227},{"_238":1223,"_658":1294,"_659":20,"_109":1225},{"_238":1071,"_658":1224,"_659":20,"_109":1240},{"_238":1073,"_658":1224,"_659":20,"_109":1291},{"_238":1238,"_658":1233,"_659":20,"_109":1239},{"_238":1292,"_658":871,"_659":20,"_109":1293},{"_238":1073,"_658":897,"_659":20,"_109":1291},{"_238":1289,"_658":773,"_659":20,"_109":1290},{"_238":1065,"_658":897,"_659":20,"_109":1284},{"_238":1261,"_658":876,"_659":73,"_109":653},{"_238":1059,"_658":1288,"_659":20,"_109":1227},{"_238":1286,"_658":1226,"_659":73,"_109":1287},{"_238":1250,"_658":773,"_659":20,"_109":1241},{"_238":1223,"_658":773,"_659":20,"_109":1225},{"_238":1285,"_658":785,"_659":20,"_109":653},{"_238":1065,"_658":773,"_659":20,"_109":1284},{"_238":1073,"_658":1218,"_659":20,"_109":1283},{"_238":1281,"_658":880,"_659":73,"_109":1282},{"_238":1279,"_658":773,"_659":73,"_109":1280},{"_238":1278,"_658":773,"_659":20,"_109":653},{"_238":1276,"_658":1224,"_659":20,"_109":1277},{"_238":1275,"_658":785,"_659":20,"_109":653},{"_238":1266,"_658":1274,"_659":73,"_109":1268},{"_238":1272,"_658":876,"_659":20,"_109":1273},{"_238":1270,"_658":961,"_659":20,"_109":1271},{"_238":1269,"_658":779,"_659":20,"_109":653},{"_238":1266,"_658":1267,"_659":73,"_109":1268},{"_238":1264,"_658":1233,"_659":73,"_109":1265},{"_238":1262,"_658":773,"_659":20,"_109":1263},{"_238":1261,"_658":776,"_659":73,"_109":653},{"_238":1232,"_658":1236,"_659":20,"_109":1234},{"_238":1064,"_658":1251,"_659":20,"_109":1241},{"_238":1259,"_658":776,"_659":20,"_109":1260},{"_238":1257,"_658":773,"_659":20,"_109":1258},{"_238":1255,"_658":773,"_659":20,"_109":1256},{"_238":1254,"_658":773,"_659":20,"_109":653},{"_238":1096,"_658":776,"_659":20,"_109":1253},{"_238":1250,"_658":897,"_659":20,"_109":1241},{"_238":1053,"_658":868,"_659":20,"_109":1252},{"_238":1250,"_658":1251,"_659":20,"_109":1241},{"_238":1249,"_658":785,"_659":20,"_109":653},{"_238":1248,"_658":773,"_659":20,"_109":1221},{"_238":1246,"_658":1233,"_659":20,"_109":1247},{"_238":1059,"_658":1244,"_659":20,"_109":1245},{"_238":1243,"_658":785,"_659":20,"_109":653},{"_238":1242,"_658":773,"_659":20,"_109":653},{"_238":1235,"_658":1233,"_659":73,"_109":1237},{"_238":1064,"_658":897,"_659":20,"_109":1241},{"_238":1071,"_658":897,"_659":20,"_109":1240},{"_238":1238,"_658":1236,"_659":20,"_109":1239},{"_238":1235,"_658":1236,"_659":73,"_109":1237},{"_238":1232,"_658":1233,"_659":20,"_109":1234},{"_238":1230,"_658":1226,"_659":73,"_109":1231},{"_238":1228,"_658":883,"_659":20,"_109":1229},{"_238":1060,"_658":1226,"_659":20,"_109":1227},{"_238":1223,"_658":1224,"_659":20,"_109":1225},{"_238":1053,"_658":875,"_659":20,"_109":1222},{"_238":1220,"_658":773,"_659":20,"_109":1221},{"_238":1071,"_658":1218,"_659":20,"_109":1219},{"_238":1216,"_658":773,"_659":20,"_109":1217},"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.

",{"_997":1000,"_900":44,"_808":44,"_629":44,"_619":44},"hasConsented","gpcApplied","showBanner","geolocation",{"_1322":1323,"_1324":1325},"country","US","state","OR","query","categories",[1345,1346,1347,1348,1349],"optionUrls",[1333,1334,1335,1336,1337,1338,1339,1340,1341,1342,1343,1344],"total",12,"/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/-",{"_442":1423,"_1351":1424,"_1353":44},{"_442":1388,"_1351":1389,"_1353":42},{"_442":443,"_1351":1372,"_1353":1373},{"_442":1364,"_1351":1365,"_1353":1366},{"_442":1350,"_1351":1352,"_1353":1354},"helm-charts","items",[1355,1356],"startIndex",10,{"_442":1350,"_449":1361,"_103":106,"_105":-7,"_444":1362,"_452":453,"_484":1363,"_1360":1343},{"_442":1350,"_449":1357,"_103":106,"_105":-7,"_444":1358,"_452":453,"_484":1359,"_1360":1344},"NVIDIA NIM Operator","k8s-nim-operator",[],"url","GPU Operator","gpu-operator",[],"resources",[1367],9,{"_442":1364,"_449":1368,"_103":106,"_105":1369,"_444":1370,"_452":453,"_484":1371,"_1360":1342},"Riva Skills Quick Start","riva","riva_quickstart",[],[1374,1375,1376],6,{"_442":443,"_449":1384,"_103":106,"_105":1385,"_444":1386,"_452":453,"_484":1387,"_1360":1339},{"_442":443,"_449":1381,"_103":106,"_105":1378,"_444":1382,"_452":453,"_484":1383,"_1360":1340},{"_442":443,"_449":1377,"_103":106,"_105":1378,"_444":1379,"_452":453,"_484":1380,"_1360":1341},"TrafficCamNet","tao","trafficcamnet",[],"PeopleNet","peoplenet",[],"StyleGAN3 pretrained models","research","stylegan3",[],"collections",[1390,1391,1392],{"_442":1388,"_449":1420,"_103":106,"_105":1394,"_444":1421,"_452":453,"_484":1422,"_1360":1336},{"_442":1388,"_449":1417,"_103":106,"_105":-7,"_444":1418,"_452":453,"_484":1419,"_1360":1337},{"_442":1388,"_449":1393,"_103":106,"_105":1394,"_444":1395,"_452":453,"_484":1396,"_1360":1338},"Omniverse Kit App Streaming","omniverse","kit-appstreaming-collection",[1397,1398,1399],{"_238":1411,"_449":1412,"_446":1413,"_457":1414,"_1404":1415},{"_238":1394,"_449":1407,"_446":1408,"_457":1403,"_1404":1409},{"_238":1400,"_449":1401,"_446":1402,"_457":1403,"_1404":1405},"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",[1406],"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.",[1406,1410],"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",[1406,1410,1416],"NGC_CONSOLE","DeepStream SDK","deepstream_sdk",[],"Omniverse Kit (FB)","kit",[1397,1398,1399],"containers",[1425,1426,1427],{"_442":1423,"_449":1434,"_103":104,"_105":1435,"_444":1436,"_452":453,"_484":1437,"_1360":1333},{"_442":1423,"_449":1431,"_103":104,"_105":106,"_444":1432,"_452":453,"_484":1433,"_1360":1334},{"_442":1423,"_449":1428,"_103":106,"_105":-7,"_444":1429,"_452":453,"_484":1430,"_1360":1335},"PyTorch","pytorch",[],"Llama-3.1-Nemotron-70B-Instruct","llama-3.1-nemotron-70b-instruct",[],"DeepSeek-R1","deepseek-ai","deepseek-r1",[],"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",{"_1467":1468,"_1469":565,"_1470":20,"_1471":1472,"_1351":1473,"_1474":1475,"_1476":1468,"_1477":20,"_1478":20,"_1479":20,"_1480":1481},"appName","ngc-catalog","appOrigin","hideOnPrivateOrg","icon","ObjectsModules2",[1482,1483,1484,1485,1486,1487,1488],"menuIcon","editor-swatches","productName","requiresAllScopes","requiresLogin","requiresOrgLevelScopesOnly","title","Catalog",{"_1480":1510,"_35":1511,"_1474":1512},{"_1503":1504,"_1480":1505,"_35":1506,"_1476":1411},{"_1480":1501,"_35":1502,"_1474":1475},{"_1480":1498,"_35":1499,"_1474":1500},{"_1480":1495,"_35":1496,"_1474":1497},{"_1480":1492,"_35":1493,"_1474":1494},{"_1480":1489,"_35":1490,"_1474":1491},"Resources","/resources","misc-library","Models","/models","misc-neural-network","Helm Charts","/helm-charts","misc-ship-wheel","Containers","/containers","3d-cube","Collections","/collections","access",[1507],"Enterprise Catalog","/enterprise",[1508,1509],"NVAIE","READ","Explore Catalog","/","common-home","urls",{"_1517":1518,"_1519":561,"_1520":563,"_1521":1522,"_1523":1524,"_1525":1526},{"_1516":20},"loggedIn","build","https://build.nvidia.com","ngcApi","ngcHelm","ngcRegistry","https://registry.ngc.nvidia.com","orgUi","https://org.ngc.nvidia.com","ngcUi","https://ngc.nvidia.com"]