NVIDIA Nemotron 3.5 Super VL

120B total / 12B active vision-language checkpoint with up to 1M combined context; official project and card dates differ.

Practical guide
Access choices, version boundaries and evaluation limits.
Official model source
Verify the exact checkpoint or provider contract.
Official model source
Verify the exact checkpoint or provider contract.

Model Specificationsnvidia

Documented examples: 4 × H100 or 2 × GB200; cache and concurrency requirements are additional.

Downloadable weights; no universal hosted token rate. Hardware/runtime costs remain.

Scoped Terms

🚀Function Overview

Primary-source contract checked October 21, 2026. No independent model request or hardware test was executed.

Key Features

Text, image and video input; text output.
Official serving examples use nemotron-3.5-super-vl; this is not a universal hosted API ID.
Model-card hardware examples: four H100 or two GB200 GPUs.
OpenMDW-1.1 governs use; commercial and non-commercial uses are described in the card.
Project collection dated October 14; the Hugging Face card dates this release October 15.

Use Cases

•Compare the exact checkpoint and provider contract
•Evaluate a permitted task with a verifiable reference

⚙️Input Parameters

model

string

nvidia/NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF16 — follow the selected official runtime or provider interface; this is not a universal hosted alias.

task_input

provider-specific

Use the documented input schema; see official sources for supported modalities and request fields.

Quick Actions

Technical Specifications

Hardware Type: Documented examples: 4 × H100 or 2 × GB200; cache and concurrency requirements are additional.
Commercial Use: OpenMDW-1.1; read its terms for the selected checkpoint.
Pricing: Downloadable weights; no universal hosted token rate. Hardware/runtime costs remain.