Nemotron 3.5 Super VL 120B A12B NVFP4
NVIDIA Nemotron 3.5 Super VL is a multimodal Mixture-of-Experts model with 120B total parameters, 12B active per token, and up to one million tokens of context. This NVFP4 version is built for visual reasoning, document understanding, and agentic workflows across images, video, and text.
Model
Nemotron 3.5 Super VL pairs a hybrid Mamba-2, Mixture-of-Experts, and attention language backbone with NVIDIA’s C-RADIO v4-H vision encoder. It processes images, sampled video frames, and text together for document and chart understanding, GUI understanding, video question answering, and multi-step agent workflows.
The model supports up to one million tokens of context, configurable thinking, and built-in Multi-Token Prediction. This version uses NVFP4 weights, while the vision encoder and projector remain in BF16. Full details are in NVIDIA’s announcement.
Dedicated Endpoint
DocsA Dedicated Endpoint is your own deployment of Nemotron 3.5 Super VL 120B A12B NVFP4, on an inference stack that Modal tunes and autoscales. Bring fine-tuned weights if you have them. Compute is billed by the second, and only while it runs.
modal endpoint create --model nvidia/NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-NVFP4