The LLM benchmark aggregator.

NVIDIA Nemotron 3 Nano 4B

NVIDIA · 4B

Strongest in Knowledge work (#152 of 177, on 1 of its 2 boards), weakest in Agents (#233 of 267). Above par in 0 of 10 scopes. Among the models it meets almost everywhere, it finishes behind Claude Sonnet 5 and GPT-5.6 Luna and ahead of G9v3 3B and Nemotron 3 Ultra.

§ 1 · Profile

What it is good at

Bars run from 50 — the average of the models each source lists — so right of the line is above par. The middle column is the gap to whoever leads that scope.

Agents41−27.1#233/2671/5

Knowledge work36.5−37.1#152/1771/2

Tool use46.2−27.8—/81✱0/1

Reasoning47.4−19.5—/178✱0/4

Mathematics47.4−18.3—/140✱0/2

Science41.9−21.6—/122✱0/1

Expert reasoning41.6−19.3—/0✱✱

Coding45.4−23.7—/164✱0/5

Agentic coding44.6−24.1—/157✱0/4

Code generation37.3−32.7—/77✱0/2

§ 2 · Head to head

What it beats, and what beats it

The same models turn up scope after scope. Counted once: where both were placed, who finished higher. Bars run right for NVIDIA Nemotron 3 Nano 4B, left for the other.

GPT-4owon 6 · lost 0

Llama 4 Scout Instructwon 6 · lost 0

Qwen3.5 0.8Bwon 6 · lost 2

DeepSeek V3won 6 · lost 2

GPT-4.1 Nanowon 6 · lost 2

GPT-4o Miniwon 6 · lost 2

Mistral Medium 3.5won 5 · lost 2

Command Awon 5 · lost 2

Grok Build 0.1won 5 · lost 2

Mercury 2.5won 4 · lost 2

Llama 4 Maverick Instructwon 5 · lost 3

Grok 4.3won 5 · lost 3

Mistral Large 3won 4 · lost 3

Qwen3.6 27Bwon 4 · lost 3

Nemotron 3 Ultra 550B A55Bwon 4 · lost 3

Nemotron 3 Ultrawon 5 · lost 4

G9v3 3Bwon 5 · lost 4

Claude 3.5 Sonnetwon 3 · lost 3

K2 Horizon 0.9Bwon 4 · lost 4

Gemini 3.5 Flash Litewon 4 · lost 4

Inklingwon 5 · lost 5

MiniMax M3won 5 · lost 5

Claude 3.5 Haikuwon 3 · lost 4

O3 Miniwon 3 · lost 5

GPT-5.4 Nanowon 3 · lost 5

GPT-5.4 Miniwon 3 · lost 5

Qwen3.7 Maxwon 3 · lost 5

Claude Opus 4won 2 · lost 4

Claude 3.7 Sonnetwon 2 · lost 4

Ling 3.0 Flashwon 2 · lost 4

Granite 4.2 30Bwon 3 · lost 6

DeepSeek R1won 2 · lost 5

Gemma 4 E4Bwon 2 · lost 5

Gemini 3.1 Flash Litewon 2 · lost 5

OpenBMB 1Bwon 2 · lost 5

GPT OSS 120Bwon 2 · lost 5

Inkling Smallwon 2 · lost 5

MiniMax M2.7won 2 · lost 5

Mimo V2.5won 2 · lost 5

Kimi K2.7 Codewon 2 · lost 5

Mimo V2.5 Prowon 2 · lost 5

Kimi K2.5won 2 · lost 5

GPT-4.1won 2 · lost 5

Grok 4.1 Fastwon 2 · lost 5

GPT-4.1 Miniwon 2 · lost 6

GPT-5 Miniwon 2 · lost 6

Qwen3.6 Pluswon 2 · lost 6

Claude Haiku 4.5won 2 · lost 6

Gemini 2.5 Flashwon 2 · lost 6

GLM-4.6won 2 · lost 6

GLM-5.3 Flashwon 2 · lost 6

Granite 4.2 3Bwon 2 · lost 8

Qwen3 235B A22B Instructwon 1 · lost 6

Kimi K2won 1 · lost 6

DeepSeek V4 Flash Visionwon 1 · lost 6

Qwen3.8 Flash Nextwon 1 · lost 6

Nemotron 3 Superwon 1 · lost 7

G9v3 39A5Bwon 1 · lost 7

Gemini 3.5 Flashwon 1 · lost 7

Grok 4.5won 1 · lost 7

Gemini 3.6 Flashwon 1 · lost 7

Kimi K2.6won 1 · lost 7

Muse Spark 1.1won 1 · lost 7

Claude Opus 4.5won 1 · lost 7

Gemma 4 31Bwon 1 · lost 8

Qwen3.5 9Bwon 1 · lost 8

Qwen3.6 35B A3Bwon 1 · lost 8

Qwen3.5 4Bwon 1 · lost 8

DeepSeek V4 Flashwon 1 · lost 8

Qwen3.8 27Bwon 1 · lost 9

Claude Opus 4.6won 0 · lost 6

Grok 3 Miniwon 0 · lost 6

Mimo V2.6 Prowon 0 · lost 6

Muse Sparkwon 0 · lost 7

GPT-5.1won 0 · lost 7

GPT-5won 0 · lost 7

GLM-4.7won 0 · lost 7

GLM-5.1won 0 · lost 7

GPT-6 Lunawon 0 · lost 7

Grok 4won 0 · lost 7

Muse Spark 1.2won 0 · lost 7

GPT-6 Solwon 0 · lost 7

Gemini 3 Prowon 0 · lost 7

GPT-6 Astrawon 0 · lost 7

Muse Spark 1.3won 0 · lost 7

Grok 4.7won 0 · lost 7

Claude Opus 5.5won 0 · lost 7

Qwen3.5 2Bwon 0 · lost 8

Gemini 2.5 Prowon 0 · lost 8

Gemma 4 12Bwon 0 · lost 8

Gemini 3.1 Prowon 0 · lost 8

Claude Sonnet 4won 0 · lost 8

K2 Horizon 32Bwon 0 · lost 8

Muse Glimmer-30Bwon 0 · lost 8

O3won 0 · lost 8

DeepSeek V3.2won 0 · lost 8

GPT-5.4won 0 · lost 8

O4 Miniwon 0 · lost 8

GPT-5.2won 0 · lost 8

Gemini 3.7 Flashwon 0 · lost 8

GPT-5.5won 0 · lost 8

Claude Sonnet 4.5won 0 · lost 8

Claude Sonnet 4.6won 0 · lost 8

Claude Opus 4.7won 0 · lost 8

Gemini 3.8 Flashwon 0 · lost 8

Claude Opus 4.8won 0 · lost 8

Grok 4.6won 0 · lost 8

Qwen3.8 Maxwon 0 · lost 8

Claude Fable 5.1won 0 · lost 8

Claude Fable 5won 0 · lost 8

Granite 4.2 8Bwon 0 · lost 9

K2 Horizon 7Bwon 0 · lost 9

K2 Horizon MoVA 36B A4Bwon 0 · lost 9

K2 Horizon 375B A23Bwon 0 · lost 9

DeepSeek V4 Prowon 0 · lost 9

DeepSeek V4.1 Flashwon 0 · lost 9

GPT-5.6 Solwon 0 · lost 9

Kimi K3won 0 · lost 9

GLM-5.3won 0 · lost 9

Claude Opus 5won 0 · lost 9

K2 Horizon 3.7Bwon 0 · lost 10

GLM-5.2won 0 · lost 10

GPT-5.6 Terrawon 0 · lost 10

GPT-5.6 Lunawon 0 · lost 10

Claude Sonnet 5won 0 · lost 10

§ 3 · Sources

Where the numbers come from

2 publications, 7 figures. Every one links to the page it was read from.

GDPval-AA ↗1 measureread 2026-09-27

GDPval-AA -74

✱✱ IFM — K2 Horizon launch ↗the model’s own publisher6 measuresread 2026-09-27

Terminal-Bench 2.1 3.7%SciCode 16.4%GPQA Diamond 51.3%HMMT Feb 2026 34.7%HLE 4.9%BFCL v4 36.8%

Badge

[![PublicAI Index](https://publicai.io/model-index/badge?model=nvidia-nemotron-3-nano-4b)](https://publicai.io/model-index/m/nvidia-nemotron-3-nano-4b)