[whatmodel.sh](../../README.md) / GPT-5 nano
# GPT-5 nano
OpenAI · released 2025-08 · closed weights
**score 61.5** ±0.8 · #188 of 732 · from 26 benchmark results across 13 sources
Score within the [LMArena (text)](https://whatmodel.sh/?set=lmarena-text&score=set) comparison set only.
## What it's good at
● marks GPT-5 nano among every model on that task or test, worst to best.
task vs the others rank counts tests
agent tasks ▁▁▃▆●▄▅▅▄▄▂▁ #184 of 313 19% 6
chat & writing ▁▁▁▃▆▄▆█●▇▇▃ #180 of 402* 6.3% 1
math ▁▁▃▅█▆▄●▄▂▁▁ #136 of 431 33% 7
vision & documents ▁▁▂▂▁▃●▂▆▇█▄ #99 of 153* 6.1% 2
agent tasks · #184 of 313 · 19% of its score
Beats [o1 (2024-12-17)](o1-2024-12-17.md), [o3-mini](o3-mini.md), [DeepSeek R1 0528](deepseek-r1-0528.md) here, though they rank higher overall. Trails [SmolLM3 3B](smollm3-3b.md), [Jamba 1.5 Mini](jamba-1-5-mini.md), [Jamba 1.5 Large](jamba-1-5-large.md) here, though they rank lower overall.
test result vs the others in its run difficulty counts why
BrowserGym WorkArena-L1 (AgentLab… 40.6% █···●▄▄·▄█·█ beats 3/10 hard · avg 28% 3.9% at its level
Terminal-Bench 2.0 with Terminus … 7.9% ▃··●·▃▃██▅▃▅ beats 1/14 very hard · avg 6.1% 3.8% sharp
BrowserGym WorkArena-L2 (AgentLab… 3.4% ▃▅●·▅▃▃··█·▃ beats 3/11 very hard · avg 3.6% 3.6% sharp · noisy
BFCL V4 (December 2025 archive) 51.4% ▁▁▁▂██▅▅●▂▂▂ beats 59/77 hard · avg 31% 3.0% at its level
τ²-bench Telecom (maker) 35.5% ▃·●▅▃▃··█▃▃▅ beats 2/13 hard · avg 16% 2.9% sharp · at its level · noisy
BrowserGym MiniWoB (AgentLab Gene… 64.8% ▅····▅●▃▃·▃█ beats 5/11 medium · avg 53% 2.3% blunt · at its level
chat & writing · #180 of 402 · 6.3% of its score
Beats [Seed-OSS 36B Instruct](seed-oss-36b.md) here, though they rank higher overall. Trails [LFM2.5 1.2B Instruct](lfm2-5-1-2b-instruct.md), [LFM2.5 1.2B Thinking](lfm2-5-1-2b-thinking.md), [LFM2.5 8B-A1B](lfm2-5-8b-a1b.md) here, though they rank lower overall.
test result vs the others in its run difficulty counts why
LMArena (text) 1337.7 ▁▁▁▂▃▃▃▆●▇█▃ beats 176/385 avg 1260 6.3% blunt
math · #136 of 431 · 33% of its score
Beats [Claude Sonnet 4.6](claude-sonnet-4-6.md), [MiniMax-M3](minimax-m3.md), [MiMo-V2.5 (310B-A15B)](mimo-v2-5.md) here, though they rank higher overall. Trails [Olmo 3.1 32B Think](olmo-3-1-32b-think.md), [Qwen3-VL 30B-A3B Thinking](qwen3-vl-30b-a3b-thinking.md), [Granite 4.2 8B](granite-4-2-8b.md) here, though they rank lower overall.
test result vs the others in its run difficulty counts why
FrontierMath tiers 1–3 v2 20.0% ▂··▁▁●▄▄█▆▄▃ beats 14/86 very hard · avg 1.2% 13% sharp
MATH level 5 94.9% ▂▂▅▆▄█▆▃▅▂●▄ beats 82/92 hard · avg 28% 8.4% sharp
OTIS mock AIME 2024–2025 81.1% ▃▃▄▃▃▃▇●█▃▂▅ beats 116/197 very hard · avg 7.2% 5.1% sharp · at its level · noisy
MATH-Vision 62.2% ▁▁·▁··▄▁●▇▇█ beats 7/24 medium · avg 47% 1.9% at its level
MathNet-Solve-Test micro-average … 42.2% ▃▃···●▃·▃·▃▃ beats 4/8 very hard · avg 15% 1.9% at its level
FrontierMath tier 4 v2 2.4% ▁●▁▂█▅▂▃▂▁▁▁ beats 7/67 very hard · avg 0.1% 1.8% sharp · too hard for it
FormalProofBench (200 graduate-le… 12.5% ▃▅·▃··●▃█··▃ beats 4/10 very hard · avg 3.1% 0.7% too hard for it · noisy
vision & documents · #99 of 153 · 6.1% of its score
Beats [DeepSeek V4 Flash Vision (Exp)](deepseek-v4-flash-vision-exp.md) here, though they rank higher overall. Trails [GPT-4o (2024-11-20)](gpt-4o-2024-11-20.md), [Qwen3-VL 30B-A3B Thinking](qwen3-vl-30b-a3b-thinking.md), [Mistral Small 4 (119B-A6.5B)](mistral-small-4.md) here, though they rank lower overall.
test result vs the others in its run difficulty counts why
Arena vision (style controlled) 1147.4 ▁▁▂▂▂▂▃●▅▅█▇ beats 47/152 avg 1061 4.2% blunt
MATH-Vision 62.2% ▁▁·▁··▄▁●▇▇█ beats 7/24 medium · avg 47% 1.9% at its level
other tests (11)
test result vs the others in its run difficulty counts why
GPQA Diamond 69.4% ▁▁▁▁▂▄▆▄●▇█▅ beats 104/211 hard · avg 39% 11% sharp · at its level
CLIcK Korean (Daekeun Kim) 72.0% ▆▄▄▄●▄▂·▄█·▄ beats 9/23 medium · avg 56% 7.5% typical
KMMLU-HARD (Daekeun Kim, zero-sho… 51.7% █···▁▁▁▂●·▆▂ beats 13/23 very hard · avg 29% 6.6% sharp
HAE-RAE 1.0 (Daekeun Kim) 78.6% ▄▄▄▂▂●▂▄▂█▂▄ beats 10/22 medium · avg 61% 6.0% typical
SimpleQA Verified 11.7% ▁▁●▂▃▃▆▆█▄▂▅ beats 3/81 very hard · avg 8.0% 1.6% typical
Chess Puzzles (Epoch) 27.0% █▁▁▁▁▂▃●▃▂▁▁ beats 100/137 very hard · avg 5.8% 1.4% sharp · too hard for it · noisy
SimpleQA (maker) 11.0% ▄▄▄▆█●███·▂█ beats 13/32 very hard · avg 9.2% 1.2% noisy
Mystery Game Puzzles (Epoch) 8.0% ●··▂▃▇▄█▂▄▂▁ beats 8/77 very hard · avg 9.4% 1.1% sharp · too hard for it · noisy
Open PT-LLM Leaderboard average (… 79.8% ▃▁▁▁▁▁▁▁▂█●▁ beats 257/276 medium · avg 72% 1.0% noisy
Vectara hallucination leaderboard… 95.3% ▁▁▂▃▄▄●▆▃▂▂▁ beats 69/140 very easy · avg 94% 0.2% blunt · maxed out · noisy
Vectara hallucination leaderboard… 89.5% ▁▂▁▂▅●▂▃▄▃▁▁ beats 39/98 very easy · avg 90% 0.0% blunt
source results share of the score
Epoch AI 8 43% ▇▇▇▇▇
Daekeun Kim Korean LLM benchmark runs 3 20% ▇▇
LMArena leaderboard dataset 2 10% ▇
ServiceNow BrowserGym leaderboard 3 10% ▇
Terminal-Bench: Benchmarking Agents on … 1 4% ▇
Berkeley Function Calling Leaderboard r… 1 3% ▇
OpenAI vendor-reported benchmark scores… 1 3% ▇
Qwen/Qwen3.5-4B model card 1 2% ▇
MathNet: a Global Multimodal Benchmark … 1 2% ▇
OpenAI (submitted / maker-reported) 1 1% ▇
Open Portuguese LLM Leaderboard 1 1% ▇
FormalProofBench 1 1% ▇
Vectara hallucination leaderboard 2 0% ▇
prior: pull toward the middle 0% ▇
weighted average = 61.5
Licences: Epoch AI ([CC BY 4.0](https://creativecommons.org/licenses/by/4.0/)) · Daekeun Kim Korean LLM benchmark runs ([MIT](../../licenses/korean-personal-benchmark-runs-MIT.txt)) · LMArena leaderboard dataset ([CC BY 4.0](https://creativecommons.org/licenses/by/4.0/)) · ServiceNow BrowserGym leaderboard ([MIT](../../licenses/hf-leaderboard-spaces-MIT.txt)) · Terminal-Bench: Benchmarking Agents on Hard ([CC BY 4.0](../../licenses/paper-2601.11868v1-CC-BY-4.0.txt)) · Berkeley Function Calling Leaderboard result archive ([Apache-2.0](../../licenses/bfcl-result-archive-v4-Apache-2.0.txt)) · Qwen/Qwen3.5-4B model card ([apache-2.0](../../licenses/Qwen--Qwen3.5-4B.txt)) · MathNet: a Global Multimodal Benchmark for Mathematical Reasoning and Retrieval ([CC BY 4.0](../../licenses/paper-2604.18584v2-CC-BY-4.0.txt)) · Open Portuguese LLM Leaderboard ([Apache-2.0](../../licenses/hf-leaderboard-spaces-Apache-2.0.txt)) · FormalProofBench ([CC-BY-4.0](https://creativecommons.org/licenses/by/4.0/)) · Vectara hallucination leaderboard ([Apache-2.0](https://www.apache.org/licenses/LICENSE-2.0))
## How to run it
### Through an API
api (OpenAI) · $0.14/M blended · $0.05 in / $0.4 out per 1M tokens · [platform.openai.com](https://platform.openai.com/docs/models)
[price source](https://github.com/anomalyco/models.dev/blob/a668908cfe7073faa0363699486fc41a3ff52412/providers/openai/models/gpt-5-nano.toml): Standard base token rate from models.dev. Provider-specific tiers, long-context surcharges, caching and promotions may differ. · [context source](https://github.com/anomalyco/models.dev/blob/a668908cfe7073faa0363699486fc41a3ff52412/models/openai/gpt-5-nano.toml)
[interactive version](https://whatmodel.sh/model/gpt-5-nano) · [JSON](https://whatmodel.sh/api?format=json&all=1)