# Nemotron 3.5 Super VL



NVIDIA



An open multimodal model from NVIDIA that mixes Mamba-2, MoE and attention layers. It targets agent work such as coding, web research, document and video understanding, and can be fine-tuned with NeMo.


## Availability

Open weights on Hugging Face and ModelScope \(BF16, NVFP4; OpenMDW-1.1 license\); hosted on OpenRouter
- Model released: Oct 15, 2026 · Date only
- Review updated: Oct 16, 2026 08:05 Fri · Beijing Time (UTC+8)

## Sources

- Source archive: https://developer.nvidia.com/blog/nvidia-nemotron-3-5-super-vl-powers-specialized-agents-with-multimodal-reasoning/
- Source archive: https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF16
- Source excerpt: https://x.com/OpenRouter/status/2110755137023246567

Canonical entry: https://www.ailyre.com/source/?id=nvidia-nemotron-3-5-super-vl&lang=en