DownloadNVIDIA Nemotron 3.5 Super VL 120B-A12B NVFP4 is an open-weight, quantized vision-language model designed for multimodal reasoning, coding, document understanding, and autonomous AI agents. It features a hybrid Mamba-2–Transformer Mixture-of-Experts architecture with 120B total parameters and approximately 12B activated per token. The model integrates a C-RADIO vision encoder to process text, images, and video, while supporting context windows up to one million tokens. NVFP4 mixed-precision quantization reduces memory requirements for efficient deployment on NVIDIA accelerators. Native Multi-Token Prediction supports faster generation through speculative decoding, while configurable reasoning modes balance response speed and analytical depth. Designed for enterprise AI, visual analysis, software engineering, and tool-assisted workflows, it supports deployment through modern inference frameworks.
Features
- 120B total parameters with 12B activated per token
- Hybrid Mamba-2–Transformer Mixture-of-Experts architecture
- 512 routed experts with a C-RADIO vision encoder
- Native text, image, and video understanding
- Up to 1M-token context window for extended workflows
- NVFP4 quantization with quantization-aware distillation
- Configurable thinking modes and Multi-Token Prediction decoding
- Deployment through vLLM, SGLang, and TensorRT-LLM
Project Samples
