Nemotron 3 Nano Omni (free)
NVIDIA · released Apr 28, 2026
NVIDIA Nemotron™ 3 Nano Omni is a 30B-A3B open multimodal model designed to function as a perception and context sub-agent in enterprise agent systems. It accepts text, image, video, and...
Specification
- Context window
- 256K
- Max output
- 66K
- Knowledge cutoff
- Not stated
- Parameters
- 33B
- Licence
- other
- Serving providers
- 1
- Moderated
- No
- Uptime
- 0.0%
Intelligence
10.3
43th percentile
Coding
13.8
Coding Index
Agentic
—
Agentic Index
Output speed
26 t/s
Median across providers
Latency
10.58s
Time to first token
Cost per task
Free
Estimated
Benchmarks
Where the score comes from
The Intelligence Index is a composite. These are the underlying evaluations this model was actually measured on.
Evaluation scores
Percentage correct · higher is better
- IFBench63.2%
- GPQA Diamond46.9%
- τ²-bench (Telecom)45.3%
- AA-LCR (long context)39.7%
- Terminal-Bench Hard8.3%
- Humanity's Last Exam4.8%
An evaluation missing from this list was not run for this model — it is not a zero.
View as table
| Evaluation | Score |
|---|---|
| IFBench | 63.2% |
| GPQA Diamond | 46.9% |
| τ²-bench (Telecom) | 45.3% |
| AA-LCR (long context) | 39.7% |
| Terminal-Bench Hard | 8.3% |
| Humanity's Last Exam | 4.8% |
Against its peers
Intelligence Index · this model highlighted, nearest peers in grey
- Trinity Large ThinkingArcee AI10.8
- Mercury 2Inception10.8
- Nemotron 3 Nano Omni (free)NVIDIA10.3
- GPT-4.1 MiniOpenAI10.2
- North Mini Code (free)Cohere9.9
- 9.9
- Gemini 2.5 FlashGoogle9.9
- R1 Distill Llama 70BDeepSeek9.8
Peers are the models sitting closest on the Intelligence Index — the set you would realistically choose between.
View as table
| Model | Intelligence |
|---|---|
| Trinity Large Thinking | 10.8 |
| Mercury 2 | 10.8 |
| Nemotron 3 Nano Omni (free) | 10.3 |
| GPT-4.1 Mini | 10.2 |
| North Mini Code (free) | 9.9 |
| Qwen3 VL 235B A22B Instruct | 9.9 |
| Gemini 2.5 Flash | 9.9 |
| R1 Distill Llama 70B | 9.8 |
Percentile among all indexed models
Pricing
What it costs to run
List prices per million tokens, plus what one representative task works out to.
List price
- Input / 1M tokens
- Free
- Output / 1M tokens
- Free
- Cached input / 1M
- Not offered
- Blended 3:1
- Free
One task, estimated
Free
- Input tokens
- 50,000
- Output tokens
- 80,000
- Profile
- Reasoning
Estimated from list pricing: 50K input tokens plus 80K output tokens for reasoning models (25K for non-reasoning).
Serving providers
Speed and latency figures are medians across these providers, so a widely-served model reports a blend rather than any single endpoint.