A multimodal Mixture of Experts (MoE) model containing 35B total parameters (3B activated)
Qwen3.6-35B-A3B
Description
A multimodal mixture-of-experts (MoE) model containing 35B total parameters (3B activated), 40 layers, and a context length of 262k tokens (extendable to 1.01M via YaRN), designed for agentic coding, multimodal reasoning, and long-context understanding across text, image, and video inputs. Key capabilities include thinking preservation across conversation turns, multi-token prediction, and tool calling for multi-step agent pipelines. Qwen3.6-35B-A3B was developed by the Qwen team. This model is ready for commercial or non-commercial use.
Third-Party Community Consideration
This model is not owned or developed by NVIDIA. This model has been developed and built to a third-party's requirements for this application and use case; see link to Non-NVIDIA Qwen Model Card.
License and Terms of Use
GOVERNING TERMS: Use of this container is governed by the NVIDIA Software License Agreement and Product-Specific Terms for NVIDIA AI Products. Use of the model is governed by the NVIDIA Open Model Agreement. ADDITIONAL INFORMATION: Apache 2.0.
Deployment Geography
Global
Use Case
Developers building AI agents, coding assistants, and multimodal applications that process text, images, and video, requiring advanced reasoning, agentic coding, and long-context understanding.
Get Help
NVIDIA Developer Community Forum
Get access to community knowledge base articles and support cases (https://forums.developer.nvidia.com/)
Release Date
NGC 07/15/2026 via Qwen3.6-35B-A3B on NGC
Reference(s)
Model Architecture
Architecture Type: Transformer Network Architecture: Mixture of Experts (MoE) with 256 experts (8 routed + 1 shared per token) This model was developed based on Qwen/Qwen3.6-35B-A3B. Number of model parameters: 35B
Input
Input Type(s): Text, Image, Video Input Format(s): Text: String; Image: Red, Green, Blue (RGB); Video: mp4, mov, webm Input Parameters: One-Dimensional (1D), Two Dimensional (2D), Three Dimensional (3D) Other Properties Related to Input: The model supports up to 262k tokens of context length (extensible to 1.01M via YaRN) and can operate in thinking or non-thinking modes via API flags. It accepts multimodal inputs (text, image, video) through the OpenAI-compatible chat endpoint.
Output
Output Type(s): Text Output Format: String Output Parameters: One-Dimensional (1D) Other Properties Related to Output: The model can preserve thinking traces across messages and supports configurable sampling parameters. Output length can be set up to 81,920 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(s): vLLM Supported Hardware Microarchitecture Compatibility:
- NVIDIA DGX Spark
Precision: FP8, NVFP4 Supported Operating System(s): Ubuntu 22.04, Ubuntu 24.04
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.
This AI model can be embedded as an Application Programming Interface (API) call into the software environment described above.
Model Version(s)
Qwen3.6-35B-A3B v3.6
Training, Testing, and Evaluation Datasets
Training Dataset
Data Modality:
- Text
- Image
- Video Image Training Data Size: Undisclosed Text Training Data Size: Undisclosed Video Training Data Size: Undisclosed Data Collection Method by dataset: Undisclosed Labeling Method by dataset: Undisclosed Properties: Undisclosed
Testing Dataset
Data Collection Method by dataset: Undisclosed Labeling Method by dataset: Undisclosed Properties: Undisclosed
Evaluation Dataset
Benchmark Results
Language Benchmarks
| Benchmark | Qwen3.5-27B | Gemma4-31B | Qwen3.5-35BA3B | Gemma4-26BA4B | Qwen3.6-35BA3B |
|---|---|---|---|---|---|
| SWE-bench Verified | 75.0 | 52.0 | 70.0 | 17.4 | 73.4 |
| SWE-bench Multilingual | 69.3 | 51.7 | 60.3 | 17.3 | 67.2 |
| SWE-bench Pro | 51.2 | 35.7 | 44.6 | 13.8 | 49.5 |
| Terminal-Bench 2.0 | 41.6 | 42.9 | 40.5 | 34.2 | 51.5 |
| MMLU-Pro | 86.1 | 85.2 | 85.3 | 82.6 | 85.2 |
| GPQA Diamond | 85.5 | 84.3 | 84.2 | 82.3 | 86.0 |
| HLE | 24.3 | 19.5 | 22.4 | 8.7 | 21.4 |
| LiveCodeBench v6 | 80.7 | 80.0 | 74.6 | 77.1 | 80.4 |
| AIME 2026 | 92.6 | 89.2 | 91.0 | 88.3 | 92.7 |
| C-Eval | 90.5 | 82.6 | 90.2 | 82.5 | 90.0 |
Vision Benchmarks
| Benchmark | Qwen3.5-27B | Claude-Sonnet-4.5 | Gemma4-31B | Gemma4-26BA4B | Qwen3.5-35BA3B | Qwen3.6-35BA3B |
|---|---|---|---|---|---|---|
| MMMU | 82.3 | 79.6 | 80.4 | 78.4 | 81.4 | 81.7 |
| MMMU-Pro | 75.0 | 68.4 | 76.9* | 73.8* | 75.1 | 75.3 |
| Mathvista (mini) | 87.8 | 79.8 | 79.3 | 79.4 | 86.2 | 86.4 |
| RealWorldQA | 83.7 | 70.3 | 72.3 | 72.2 | 84.1 | 85.3 |
| MMBench EN-DEV-v1.1 | 92.6 | 88.3 | 90.9 | 89.0 | 91.5 | 92.8 |
| OmniDocBench1.5 | 88.9 | 85.8 | 80.1 | 74.4 | 89.3 | 89.9 |
| VideoMMMU | 82.3 | 77.6 | 81.6 | 76.0 | 80.4 | 83.7 |
| VideoMME (w sub.) | 87.0 | 81.1 | -- | -- | 86.6 | 86.6 |
| VideoMME (w/o sub.) | 82.8 | 75.3 | -- | -- | 82.5 | 82.5 |
| RefCOCO (avg) | 90.9 | -- | -- | -- | 89.2 | 92.0 |
| ODInW13 | 41.1 | -- | -- | -- | 42.6 | 50.8 |
Data Collection Method by dataset: [Hybrid: Automated, Manually-Collected] Labeling Method by dataset: [Hybrid: Automated, Manually-Labeled] Properties: The model is evaluated on benchmarks covering agent capabilities, vision-language tasks, general visual question answering, STEM and reasoning, knowledge, and spatial intelligence.
Inference
Acceleration Engine: vLLM Precision: FP8, NVFP4 Test Hardware:
- NVIDIA DGX Spark
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. Developers should work with their internal developer team to ensure these software components meet requirements for the relevant industry and use case and address unforeseen product misuse.
Please make sure you have proper rights and permissions for all input image and video content; if image or video includes people, personal health information, or intellectual property, the image or video generated will not blur or maintain proportions of image subjects included.
Users are responsible for model inputs and outputs. Users are responsible for ensuring safe integration of this model, including implementing guardrails as well as other safety mechanisms, prior to deployment.
Please report model quality, risk, security vulnerabilities or NVIDIA AI Concerns here.