Noesion Radarnoesion.ai
Release2026-10-15

NVIDIA released Nemotron 3.5 Super VL. It is an open multimodal mixture-of-experts model. It uses 12B active parameters out of 120B total. It accepts text, images, documents, GUI screenshots and video as input. It returns text. It is built for long-running agents. It uses a hybrid Mamba-Transformer architecture with multi-token prediction. It supports a 1M-token context. Reasoning is on by default. It can be turned off or given a token budget. Typical uses include document intelligence, web research, coding validation, customer service and video understanding. It supports tool calling and structured outputs. It is served by five providers on OpenRouter.

What it unlocks

Running long agents that combine visual understanding with reasoning and tool use.

  • 12B active parameters out of 120B total
  • $0.10 per 1M input tokens
  • $0.40 per 1M output tokens
  • 1M-token context window

Sources

Send this to someone who needs it

Shares the story and its sources. Nothing about you.

What does this mean for your job?

This is the story as everyone gets it. Once a week we send you the version written for your role — what changed, why it matters for the work you actually do, and one thing to try. Free while we tune it.