Request a pilot
‹ PublicAI Index

The LLM benchmark aggregator.

GPT-5 Nano

OpenAI

Strongest in Harm refusal (#4 of 300), weakest in Human preference (#179 of 342). Above par in 14 of 24 scopes. Among the models it meets almost everywhere, it finishes behind Claude Fable 5 and Claude Fable 5.1 and ahead of Grok 3 and Kimi K2.

§ 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.

Safety59.6#11/337
Harm refusal60.7#4/300
Secure code65.9#8/274
Jailbreak resistance62.9#11/272
Safe-prompt compliance57.7#17/82
Fairness61.1#37/300
Factual grounding48.7#61/101
Toxicity avoidance54.6#121/272
Agents55.5#59/268
Tool use59.9#20/81
Core abilities53.2#60/204
General intelligence54.6#63/204
Reasoning47.1#108/178
Mathematics54#61/140
Reasoning43.3#93/140
Science41.2#96/122
Knowledge44.6#108/138
Academic knowledge43.6#99/123
Professional43.1#142/168
Finance46.1#108/151
Medical45.4#110/140
Legal35.5#143/151
Human preference51#179/342
Human preference51#179/342

§ 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 GPT-5 Nano, left for the other.

GPT-4.1 Nanowon 22 · lost 1
GPT-4o Miniwon 20 · lost 1
Gemini 1.5 Flashwon 16 · lost 1
Jamba 1.6 Miniwon 14 · lost 1
Jamba 1.6 Largewon 14 · lost 1
Command R Pluswon 16 · lost 2
Llama 4 Scout Instructwon 21 · lost 3
Mistral Smallwon 18 · lost 3
Mistral Largewon 15 · lost 4
GPT OSS 20Bwon 16 · lost 5
Gemini 1.5 Prowon 14 · lost 5
Command Awon 16 · lost 6
Claude 3.5 Haikuwon 12 · lost 5
GPT-4owon 14 · lost 6
Gemini 2.0 Flashwon 13 · lost 8
GPT-4.1 Miniwon 14 · lost 9
DeepSeek V3won 14 · lost 9
Grok 4.3won 12 · lost 9
O3 Miniwon 11 · lost 9
Kimi K2won 13 · lost 11
Grok 3won 11 · lost 10
GPT-6 Lunawon 9 · lost 9
Gemini 2.5 Flash Litewon 10 · lost 10
MiniMax M3won 11 · lost 11
Qwen3.7 Maxwon 10 · lost 11
Gemini 3.6 Flashwon 10 · lost 11
Gemma 4 31Bwon 9 · lost 10
Mistral Large 3won 8 · lost 9
Ling 3.0 Flashwon 7 · lost 8
GPT-4.1won 11 · lost 13
GPT-5.6 Lunawon 10 · lost 12
Kimi K2.6won 10 · lost 12
Mimo V2.5 Prowon 9 · lost 11
GPT-5.4 Miniwon 7 · lost 9
GPT OSS 120Bwon 10 · lost 13
DeepSeek V3.2won 10 · lost 13
DeepSeek V4 Flashwon 9 · lost 12
GLM-5.3 Flashwon 9 · lost 12
Gemini 3.5 Flashwon 9 · lost 12
GLM-4.5won 8 · lost 11
Gemini 2.5 Flashwon 10 · lost 14
DeepSeek R1won 9 · lost 13
Inklingwon 9 · lost 13
Gemini 3.5 Flash Litewon 6 · lost 9
Claude 3.7 Sonnetwon 8 · lost 12
Claude Sonnet 4.6won 8 · lost 12
GPT-5.1won 8 · lost 13
O4 Miniwon 9 · lost 15
Claude Haiku 4.5won 9 · lost 15
GLM-5.1won 7 · lost 12
GLM-4.6won 8 · lost 14
Claude Opus 4won 8 · lost 14
Kimi K2.5won 8 · lost 14
GLM-5.2won 8 · lost 14
O3won 8 · lost 15
Qwen3.6 Pluswon 5 · lost 10
Grok 4.1 Fastwon 7 · lost 14
Grok 4.5won 7 · lost 14
Qwen3.8 Maxwon 7 · lost 14
Gemini 3.8 Flashwon 7 · lost 14
Qwen3 235B A22B Instructwon 8 · lost 16
Claude Sonnet 4.5won 8 · lost 16
GPT-5.6 Terrawon 7 · lost 15
GPT-5.6 Solwon 7 · lost 15
O1won 6 · lost 13
GPT-6 Solwon 6 · lost 13
Claude Opus 4.6won 6 · lost 13
Mimo V2.6 Prowon 5 · lost 11
Claude Sonnet 4won 7 · lost 16
Grok 4won 7 · lost 16
GPT-5won 7 · lost 16
Gemini 2.5 Prowon 7 · lost 16
GPT-5 Miniwon 7 · lost 17
Muse Spark 1.1won 6 · lost 15
Kimi K2 Thinkingwon 5 · lost 13
Claude 3.5 Sonnetwon 5 · lost 13
DeepSeek V4 Prowon 6 · lost 16
GPT-5.5won 6 · lost 16
GPT-5.2won 6 · lost 17
Qwen3.8 27Bwon 4 · lost 12
Grok 4.6won 5 · lost 16
Gemini 3 Prowon 5 · lost 16
GPT-5.4 Nanowon 4 · lost 13
Gemini 3.1 Prowon 4 · lost 13
Claude Sonnet 5won 5 · lost 17
GPT-5.4won 5 · lost 17
MiniMax M2.7won 4 · lost 14
Grok 4.7won 4 · lost 14
Claude Opus 4.7won 4 · lost 18
Claude Opus 4.5won 3 · lost 15
Claude Opus 4.8won 3 · lost 18
Claude Opus 5won 3 · lost 18
Grok 4 Fastwon 2 · lost 14
Gemini 3.1 Flash Litewon 2 · lost 14
Claude Opus 5.5won 2 · lost 16
GLM-5.3won 1 · lost 14
Kimi K3won 1 · lost 14
GLM-5won 1 · lost 15
Gemini 3.7 Flashwon 0 · lost 15
Claude Fable 5.1won 0 · lost 15
Claude Fable 5won 0 · lost 15

§ 3 · Sources

Where the numbers come from

8 publications, 26 figures. Every one links to the page it was read from.

LMArena Text ↗1 measureread 2026-09-26
LMArena Text 1338
ARC-AGI-2 ↗1 measureread 2026-09-26
ARC-AGI-2 2.6%
BFCL ↗1 measureread 2026-09-26
BFCL v4 51.45%
Kagi LLM Benchmark ↗1 measureread 2026-09-26
Kagi LLM Benchmark 62.2%
Vals.ai ↗10 measuresread 2026-09-26
Vals · CaseLaw 52.63%Vals · LegalBench 50.13%Vals · TaxEval 67.38%Vals · MortgageTax 53.62%Vals · MedQA 93.26%Vals · MedCode 30.44%Vals · MedScribe 72.86%Vals · GPQA Diamond 63.38%Vals · MMLU Pro 76.07%Vals · AIME 81.18%
HELM Safety ↗5 measuresread 2026-09-26
HELM Safety · HarmBench 98.4%HELM Safety · SimpleSafetyTests 100%HELM Safety · Anthropic Red Team 99.6%HELM Safety · BBQ 97.6%HELM Safety · XSTest 97.6%
Enkrypt AI Safety Leaderboard ↗6 measuresread 2026-09-26
Enkrypt · Jailbreak risk 1.9%Enkrypt · Harmful content risk 7.2%Enkrypt · CBRN risk 2%Enkrypt · Toxicity risk 3.1%Enkrypt · Bias risk 61%Enkrypt · Insecure code risk 1.3%
Vectara · Factual consistency 89.5%

Badge

PublicAI Index badge for GPT-5 Nano[![PublicAI Index](https://publicai.io/model-index/badge?model=gpt-5-nano)](https://publicai.io/model-index/m/gpt-5-nano)