Skip to content

All Models

Rank #8 / 9

Nemotron 3 Nano Omni 30B-A3B

Nemotron 3 Nano Omni 30B-A3B is NVIDIA's Nemotron model (open weights). It is ranked 8 of 9 models on the Visual Intelligence Index.

  • NVIDIA
  • Open weights
  • Nemotron
  • Video input
  • Audio
  • Structured output
  • Tool calling
  • Streaming
  • Batch API
  • Self-hostable
Model Card
57.5%Visual Intelligence Index · Rank #8 of 9

Capability profile

Comparison Summary

Nemotron 3 Nano Omni 30B-A3B ranks 8 of 9 on the perception index. Immediately ahead is Qwen3.5-35B-A3B. Immediately behind is Molmo 2 8B. Weights are open and the model runs on your own hardware, so there is no provider list price to quote. The context window is 262K. Every score is from our own evaluation runs; list prices are as published by the provider. Strongest capability in this set is Causal Reasoning (64%). Weakest is Action / Event Understanding (56%).

Visual Intelligence
60.3%
Availability
Open weights
Input / 1M tokens
—
Output / 1M tokens
—
Context Window
262K
Parameters
30B

All benchmark scores

BenchmarkMetricScoreCoverageSetting
Video-MME v2Accuracy, no subtitles35.0%Full1 fps, 727-frame cap (254 videos thinned)
LVBenchTest accuracy54.2%Full1 fps, 727-frame cap (103 videos thinned)
Perception TestOverall accuracy70.3%Full1 fps
NExT-QAAccuracy (hard split)81.2%Full1 fps
Q-Bench VideoOverall accuracy64.9%Full1 fps
EgoSchemaAccuracy (fullset)65.8%Full1 fps
UCF101-ADAccuracy31.0%Full1 fps

Benchmark Scores

Compare reported model scores across each available benchmark or capability index.

9 of 9 models
Filter by model access

Display

Capabilities Index Scores

Compare reported model scores across each available benchmark or capability index.

9 of 9 models
Filter by model access

Display

Usability

5.0/ 5

What it's like to build against this model, scored out of five from the developer-facing capabilities in the dataset.

Reachable1.0

There is an endpoint you can call without hosting anything.

  • First-party API — supported
  • Third-party API — supported
Portable1.0

You can run it yourself, and are not tied to one vendor.

  • Open weights — supported
  • Self-hostable — supported
Multimodal input1.0

It takes the footage directly, rather than frames you extracted.

  • Video ingestion — supported
  • Audio — supported
Programmable1.0

Output you can parse, and tools it can call on its own.

  • Structured output — supported
  • Tool calling — supported
Operable1.0

Usable interactively and in bulk, not only one call at a time.

  • Streaming — supported
  • Batch API — supported

Other NVIDIA models

Other models from the NVIDIA family.

DateModelVisual Intelligence
Apr 28, 2026Nemotron 3 Nano Omni 30B-A3B60.3%