GPT-5 Nano
Aug 2025
GPT-5 NanoCompare inon
Benchmarks of GPT-5 Nano42
BenchmarkByScoreRank
Overall1
Knowledge4
Reasoning4
Math6
Science2
Coding1
Multimodal2
Writing14
Health & law3
Documents1
Legacy · saturated or no longer reported4
No benchmark matches
Rankingat high effort?
Alternatives?
ModelScoreCostPer taskWhat changes
CheaperSolar Mini 4×2.7-36%$0.37No image inputBetterGPT-6 Lunalow×2.1+45%$0.005Function calling only with none · Chat Completions+3
OpenMiMo v2.6 Prothinking×5.2+295%$0.13Pass back reasoning_content+3
CheaperLing 3.0 FlashthinkingOpen×2.5-35%$0.06 / $0.18Context: 400K → 262K+1CheaperLing 3.0 Flash FinthinkingOpen×2.4-35%$0.06 / $0.18Context: 400K → 262K+1ModelScoreCostPer taskWhat changes
Solar Mini 4×2.7-36%$0.37No image inputLing 3.0 FlashthinkingOpen×2.5-35%$0.06 / $0.18Context: 400K → 262K+1Ling 3.0 Flash FinthinkingOpen×2.4-35%$0.06 / $0.18Context: 400K → 262K+1Ling 3.0 Flash VLthinkingOpen×2.3-35%$0.06 / $0.18Context: 400K → 262KGPT OSS 120BhighOpen×2.1-56%$0.092Context: 400K → 131K+1ModelScoreCostPer taskWhat changes
GPT-6 Lunalow×2.1+45%$0.005Function calling only with none · Chat Completions+3GPT-5.6 Lunalow×2.1+227%$0.01MiMo v2.5thinkingOpen×2.9+27%$0.019
Claude Haiku 5.5low×2.8+45%$0.024Remove sampling parameters+5
Mercury 2high+80%+173%$0.039Context: 400K → 128K+1ModelScoreCostPer taskWhat changes
MiMo v2.6 ProthinkingOpen×5.2+295%$0.13Pass back reasoning_content+3
Kimi K3maxOpen×4.9+4,264%$2.00Sampling parameters are fixed+3
Step 5 PreviewthinkingOpen×4.9+936%$1.03web_search not available · Chat Completions+2
Muse Spark OpenxhighOpen×4.6+1,355%$0.84Reasoning off leaves output unchanged · Self-hosted / llama.cpp+1
DeepSeek V4.1 Flash 1017maxOpen×4.6+282%$0.26Pass back reasoning_content · Chat Completions+4ModelScoreCostPer taskWhat changes
GPT-6 Astrahigh×5.6—$1.73Monitoring can return HTTP 403+2GPT-6.1 Solhigh×5.4—$0.32Remove unsupported parameters+5GPT-5.6 Solhigh×4.8—$0.81GPT-5.5high×4.7—$1.54Prompts with >272K input tokensGPT-6 Solhigh×4.6—$0.37GPT-5.3 Codexxhigh×4.5—$1.75 / $14DeprecatedGPT-5.6 Terrahigh×4.1—$0.34Prompts with >272K input tokensGPT-5.6 Lunahigh×3.5—$0.044All OpenAI models →History
5 versionsBenchmarks
all categoriesChat & Knowledge
Humanity's Last Exam9.5#200
Arena Text · overall1,338#135
Arena Text · Chinese1,356#126
Arena Text · English1,321.5#139
Arena Text · German1,326.6#106
Arena Text · instruction following1,306.1#129
Arena Text · Japanese1,226.4#102
Arena Text · Korean1,269.3#103
Arena Text · non-English1,312.8#130
Arena Text · Polish1,313.7#104
Arena Text · Russian1,295.9#132
Arena Text · Spanish1,360.3#105
AA Omniscience Index-28.7#147
CritPt (physics)0#180
AA Intelligence Index 4.3.2≈13#167
AA Omniscience · Accuracy19.1#163
MathArena · BRUMO 202580.8#23
MathArena · CMIMC 202574.4#13
Arena Text · overall · full cohort1,319.9#134
GPQA Diamond67.6#199
AIME 2025 · AA archived83.7#34
Coding
LiveCodeBench · AA archived78.9#18
Agents & Tool Use
τ²-bench · AA36.5#134
Writing & Language
Arena Text · creative writing1,246#161
IFBench (instructions)67.6#60
Arena Text · creative writing · full cohort1,249.4#147
Vision & Documents
MMMU-Pro (vision)61#97
Arena Vision · overall1,159.5#72
FAQ
How much does GPT-5 Nano cost?›
$0.05 per 1M input tokens and $0.40 per 1M output tokens; cached input $0.005 per 1M.
Is GPT-5 Nano free to use?›
There is no free API; the cheapest paid price is $0.05 per 1M input tokens.
How good is GPT-5 Nano, and what is it best at?›
#199 of 334 in Chat & Knowledge, #117 of 239 in Coding, #132 of 242 in Agents & Tool Use, #161 of 195 in Writing & Language, #95 of 124 in Vision & Documents. Strongest at: Instructions.
Is GPT-5 Nano better than Ring Flash 2.0?›
On Humanity's Last Exam, GPT-5 Nano scores 9.5 vs 9.6 — about the same.
How can I use GPT-5 Nano?›
Through the API of Azure, OpenAI.
What is the context window of GPT-5 Nano?›
400K tokens, with up to 128K output tokens.
When was GPT-5 Nano released? Is there a newer version?›
Released 2025-08-07. It has been superseded by GPT-5.4 Nano.
Is GPT-5 Nano open source?›
No, GPT-5 Nano is a closed model available through an API. The closest open-weights option here is MiMo v2.6 Pro.
Can GPT-5 Nano read images?›
It accepts text and images, PDF files as input.
What is GPT-5 Nano?›
GPT-5 Nano is a closed language model from OpenAI, released Aug 2025.
What are the best alternatives to GPT-5 Nano?›
Solar Mini 4 (cheaper), GPT-6 Luna (better), MiMo v2.6 Pro (open weights).