Nemotron 3 Nano Omni 30B A3B

Nemotron 3 Nano Omni 30B A3B is a reasoning model from NVIDIA in the Nemotron 3 Nano Omni family. 16 benchmarks count toward its index, in 7 categories.

  • Open weights
  • Reasoning
  • Input: text
  • Released 2026-04
RankedEnough results for an index.
IndexAll categories, out of 100.38.6±4.5Median 51.0
CoverageIndex weight that has results.95%
SpeedOutput tokens per second.228/sMedian 67/s
First tokenSeconds to the first output token.9.5 sMedian 1.3 s
Input / 1MUS dollars per 1M input tokens.—Median $0.60
Output / 1MUS dollars per 1M output tokens.—Median $2.55
ContextMaximum tokens in one request.256KMedian 262K
EloLMArena rating and rank.—Median 1422

Run locally

33.02B parameters
16-bit
70.3 GB
8-bit
37.8 GB
4-bit
22.0 GB
Weights
Hugging Face (opens in a new tab) ↗

Memory need with a short context. The 4-bit copy fits in 24 GB. Compare local models

Capabilities

Index per category, out of 100.
Out of 100
AgenticMulti-step tasks with tools.2 counted33.4193 / 218
CodingCode writing and repair.1 counted37.6—
ReasoningLogic problems and puzzles.2 counted35.7187 / 227
MultimodalTasks with images and text.6 counted40.875 / 84
KnowledgeFacts and expert knowledge.3 counted38.2207 / 268
MultilingualTasks in many languages.None counted——
InstructionTasks with strict prompt rules.1 counted53.2—
MathMath problems.1 counted43.6—

Results

16 counted
BenchmarkThe benchmark name.ResultThe score from the source.PlacePlace among results on this benchmark.
RefCOCO averageMultimodal90.5%4 / 10
Artificial Analysis IFBenchInstruction63.2%59 / 146
Instruction Following BenchmarkInstructionAbove fitFar above this model's index.74.2%33 / 57
Graduate-Level Google-Proof Q&AKnowledge72.2%86 / 141
Critical Physics TasksReasoning0.0%150 / 214
GDPval-AA normalizedAgentic0.0%101 / 133
Scientific Code BenchmarkCoding32.0%29 / 35
GPQA Diamond · BenchLMKnowledge72.2%70 / 84
ScreenSpot ProMultimodal57.8%17 / 20
Artificial Analysis MMMU-ProMultimodal53.2%99 / 113
Artificial Analysis GPQA DiamondKnowledge46.9%180 / 199
CharXiv ReasoningMultimodal76.3%45 / 49
AI2D test splitMultimodal88.5%—
Video-MME without subtitleMultimodal72.2%—
LiveCodeBench v5Coding63.2%—
MMLongBench-DocMultimodal57.5%—
OSWorldAgentic47.4%—

Sources

BenchLMCC BY-NC 4.0

History

5 releases
Index38.2 → 38.6
23 Oct 202612 Nov 2026

Questions

How good is Nemotron 3 Nano Omni 30B A3B?
Its index is 38.6 out of 100. That is place 217 of 283 ranked models.
What is Nemotron 3 Nano Omni 30B A3B best at?
Knowledge: 38.2 out of 100, place 207 of 268.
How fast is Nemotron 3 Nano Omni 30B A3B?
228 output tokens per second, 9.5 s to the first token.
What is the context window of Nemotron 3 Nano Omni 30B A3B?
256K tokens.
Is Nemotron 3 Nano Omni 30B A3B open source?
Yes. Its weights are published.
When was Nemotron 3 Nano Omni 30B A3B released?
April 2026.

Badge

Nemotron 3 Nano Omni 30B A3B on Inference360: place 217 of 283, index 38.6
HTML
<a href="https://inference360.ai/models/nemotron-3-nano-omni-30b-a3b"><img src="https://inference360.ai/badges/nemotron-3-nano-omni-30b-a3b.svg" alt="Nemotron 3 Nano Omni 30B A3B on Inference360: place 217 of 283, index 38.6" height="20"></a>
Markdown
[![Nemotron 3 Nano Omni 30B A3B on Inference360: place 217 of 283, index 38.6](https://inference360.ai/badges/nemotron-3-nano-omni-30b-a3b.svg)](https://inference360.ai/models/nemotron-3-nano-omni-30b-a3b)

Compare with

DeepSeek-R138.9ZAYA1-8B38.2Qwen3 235B-A22B38.1Nemotron 3 Nano 30B32.9Nemotron 3 Ultra50.8

More from NVIDIA

All models
Cosmos3-NanoUnrankedCosmos3-SuperUnrankedNemotron 3 Ultra50.8Nemotron 3.5 ASR Streaming 0.6BUnrankedNemotron 3.5 Content SafetyUnranked