Inference
EN
NVIDIA

Nemotron 3 Nano 30B

Nemotron 3 Nano 30B is a standard model from NVIDIA. 10 benchmarks count toward its index, in 6 categories.

  • Open weights
  • Standard
  • Input: text
  • Released 2025-12
RankedEnough results for an index.
IndexAll categories, out of 100.32.9±6.6Median 51.0
CoverageIndex weight that has results.85%
SpeedOutput tokens per second.270/sMedian 67/s
First tokenSeconds to the first output token.8.3 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.32KMedian 262K
EloLMArena rating and rank.—Median 1422

Run locally

31.58B parameters
16-bit
67.3 GB
8-bit
36.2 GB
4-bit
21.1 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 counted26.8206 / 218
CodingCode writing and repair.1 counted40.5—
ReasoningLogic problems and puzzles.2 counted25.7215 / 227
MultimodalTasks with images and text.None counted——
KnowledgeFacts and expert knowledge.3 counted34.7226 / 268
MultilingualTasks in many languages.None counted——
InstructionTasks with strict prompt rules.1 counted24.1—
MathMath problems.1 counted47.4—

Results

10 counted
BenchmarkThe benchmark name.ResultThe score from the source.PlacePlace among results on this benchmark.
American Invitational Mathematics Examination 2025Math89.1%19 / 53
Massive Multitask Language Understanding ProfessionalKnowledge78.3%42 / 75
Graduate-Level Google-Proof Q&AKnowledge73.0%85 / 141
Critical Physics TasksReasoning0.0%150 / 214
LiveCodeBench v6Coding68.3%43 / 59
GDPval-AA normalizedAgentic0.0%101 / 133
τ²-Bench Tool-Agent-User EvaluationAgentic25.4%119 / 149
Artificial Analysis IFBenchInstruction37.5%117 / 146
Artificial Analysis Humanity's Last ExamKnowledge4.6%185 / 222
Artificial Analysis Long Context ReasoningReasoning10.7%198 / 217
Humanity's Last ExamKnowledge10.6%73 / 77
Artificial Analysis GPQA DiamondKnowledge39.9%191 / 199
Artificial Analysis Openness IndexKnowledge83.3%—

Sources

BenchLMCC BY-NC 4.0

History

5 releases
Index39.5 → 32.9
23 Oct 202612 Nov 2026

Questions

How good is Nemotron 3 Nano 30B?
Its index is 32.9 out of 100. That is place 244 of 283 ranked models.
What is Nemotron 3 Nano 30B best at?
Knowledge: 34.7 out of 100, place 226 of 268.
How fast is Nemotron 3 Nano 30B?
270 output tokens per second, 8.3 s to the first token.
What is the context window of Nemotron 3 Nano 30B?
32K tokens.
Is Nemotron 3 Nano 30B open source?
Yes. Its weights are published.
When was Nemotron 3 Nano 30B released?
December 2025.

Badge

Nemotron 3 Nano 30B on Inference360: place 244 of 283, index 32.9
HTML
<a href="https://inference360.ai/models/nemotron-3-nano-30b"><img src="https://inference360.ai/badges/nemotron-3-nano-30b.svg" alt="Nemotron 3 Nano 30B on Inference360: place 244 of 283, index 32.9" height="20"></a>
Markdown
[![Nemotron 3 Nano 30B on Inference360: place 244 of 283, index 32.9](https://inference360.ai/badges/nemotron-3-nano-30b.svg)](https://inference360.ai/models/nemotron-3-nano-30b)

Compare with

Granite 4.2 3B33.2Grok 4.1 Fast33.3Exaone 4.0 32B32.4Nemotron Ultra 253B35.9Nemotron 3 Nano Omni 30B A3B38.6

More from NVIDIA

All models
Nemotron 3 Nano 30B A3BUnrankedNemotron 3 Super 100BUnrankedNemotron 3 Super 120B A12BUnrankedParakeet TDT 0.6B v3UnrankedNemotron 3 Nano Omni 30B A3B38.6