available
Weights
Official model weights are downloadable from the linked model release.
Inspect this claim’s source ↗NVIDIA / FAMILY RECORD
Nemotron 3 Nano Omni
A compact sparse reasoning model that accepts video, audio, images, and text while activating about three billion parameters per token.
Checked release: Nemotron-3-Nano-Omni-30B-A3B-Reasoning
Sources checked 2026-09-19
- NVIDIA technical report publication
- · Primary source
- NVIDIA Build availability
- · Primary source
The profile describes this checkpoint, not every release carrying the family name. Developer documentation is not independent reproduction.
THE MODEL / THE TERMS
Know what you are picking up.
- Parameters
- 33B listed deployment parameters on NVIDIA Build; the same page’s architecture text says 31B A3B, while the model identifier names a 30B-A3B backbone. Accounting differs across first-party fields.
- Architecture
- Native omni-understanding sparse mixture-of-experts model in the Nemotron 3 family.
- Modalities
- text · image · audio · video
- Context
- 262K tokens on NVIDIA Build; the April 27 technical report describes a 256K maximum context
- Weights license
- NVIDIA Open Model Agreement
- Commercial use
- Commercial deployment is governed by NVIDIA’s custom Open Model Agreement; users must review its conditions.
- Memory
- No single minimum is asserted; requirements depend on precision, media length, and runtime.
- Known quantizations
- Check the official model page for supported NVIDIA deployment formats
Specifications are developer-reported. Primary release documentation ↗ · All sources and terms
NINE DIMENSIONS / NO OPENNESS SCORE
What is actually available?
Each component has its own evidence. “Not established” means this source review did not verify availability; it is not a claim that nothing exists.
available
Architecture
The official model page documents parameter activation, modalities, and configuration.
Inspect this claim’s source ↗available
Inference code
NVIDIA publishes reference deployment and inference guidance.
Inspect this claim’s source ↗partial
Training code
Nemotron and NeMo tooling is public, but complete code for this exact training run was not established.
Inspect this claim’s source ↗partial
Training recipe
NVIDIA documents the model’s construction and reasoning behavior without a complete end-to-end reproduction recipe.
Inspect this claim’s source ↗partial
Data information
Training approach information is published, but the full multimodal mixture is not established here.
Inspect this claim’s source ↗not established
Training data
No sufficiently specific public artifact was confirmed in this review.
Inspect this claim’s source ↗partial
Evaluation
Vendor evaluation results and methods are published on the official model page.
Inspect this claim’s source ↗partial
Commercial use
NVIDIA explicitly permits commercial use under its custom agreement. This is permission with conditions, not a prohibition and not an OSI certification.
Inspect this claim’s source ↗WHY THIS RELEASE MATTERS
What to explore.
- Video and audio understanding
- Reasoning
- Efficient active parameter count
- Long context
WHAT THE RECORD DOES NOT ESTABLISH
Keep the limits.
- Custom license
- Complete multimodal training corpus is not released
- Vendor benchmarks need independent reproduction
No independent reproduction of the Nano Omni benchmark suite is attached; performance statements remain vendor-reported.
CONNECTED EVIDENCE
A release becomes a starting point.
Thematic or family connections, not proof that an older application uses the selected checkpoint.
FOLLOW THE RECORD
Sources, not superlatives.
- Nemotron 3 Nano Omni technical report ↗ (report)
- Nemotron 3 Nano Omni 30B-A3B Reasoning ↗ (institution)
- Nemotron AI models ↗ (institution)
Source check: 2026-09-19. AI-assisted editorial review, with no named external expert or independent hardware validation claimed. Licenses and model cards can change. Review the linked terms for the specific artifact and intended use.
Understand open source vs open weights ↗