Model by NVIDIA

Nemotron 3.5 Super VL

Nemotron 3.5 Super VL is NVIDIA's open multimodal mixture-of-experts model, with 12B active parameters out of 120B total.

Checked 2026-11-11

Nemotron 3.5 Super VL is NVIDIA's open multimodal mixture-of-experts model, with 12B active parameters out of 120B total. It reads text, images, documents, GUI screenshots and video and answers in text. OpenRouter listed it on 15 October 2026 with a 1,048,576-token context window. The price there is $0.15 per million input tokens and $0.50 per million output, checked on 5 November 2026, which makes it one of the cheaper vision models in this batch. OpenRouter links public weights on Hugging Face. The Hugging Face repository for the 120B BF16 weights is public and marks the licence as other, with terms we have not read, so read them before commercial use. The 120B size means self-hosting needs serious GPUs, and output is text only, so it does not draw or speak. It fits document and screen-reading agents where cost matters. You reach it through OpenRouter or NVIDIA's own services. Figures come from OpenRouter's public model list and its model page. We have not run the model.

Mentioned here

  • Nemotron 3 Ultra

    Nemotron 3 Ultra is NVIDIA's open-weights model, released 2026-06-04, and its listing name shows 550 billion total parameters with 55 billion active.

    • Open weights

    NVIDIA, 262K context, $0.60 in / 1M

  • Gemma 4 31B

    Gemma 4 31B is Google's open model, released 2026-04-02 under Apache 2.0, a standard permissive licence.

    • Open weights

    Google, 262K context, $0.20 in / 1M

  • Qwen3.8 27B

    Qwen3.8 27B is an open model from Alibaba's Qwen team, released 2026-08-14 under Apache 2.0, a standard permissive licence.

    • Open weights

    Alibaba (Qwen), 1M context, $0.15 in / 1M

  • MiniMax M3

    MiniMax M3 is an open-weights model released 2026-05-31, with a 1,048,576-token context and text, image and video input.

    • Open weights

    MiniMax, 1M context, $0.30 in / 1M

  • OpenRouter

    OpenRouter routes more than 600 models through one API key.

    • Model gateway (router)
    • OpenAI compatible

  • Hugging Face Inference

    Hugging Face Inference Endpoints deploys a model from the Hub onto dedicated, autoscaling hardware.

    • Model / inference API
    • OpenAI compatible

Questions

What is the context window of Nemotron 3.5 Super VL?

1,048,576 tokens, with up to 235,929 tokens of output.

How much does Nemotron 3.5 Super VL cost?

$0.15 per million input tokens and $0.5 per million output tokens, checked 31 October 2026.

Can I run Nemotron 3.5 Super VL on my own hardware?

Yes, the weights are published under the Open-weights model licence. They are on huggingface.co.

Who makes Nemotron 3.5 Super VL?

NVIDIA makes it, and released it on 15 October 2026.

Reviews

Loading reviews…

Write a review

Your rating

Up to 1000 characters. Stars alone are enough.

Similar

  • Nemotron 3.5 Lightning

    Nemotron 3.5 Lightning is NVIDIA's small, fast open model, and it suits developers who want an agent model on a single GPU.

    • Open weights

    NVIDIA, 262K context, $0.07 in / 1M