OpenAIReasoning modelGPT-5 nano
Best configuration ranks #270 of 610 on the BenchLeader Index at 51.1 ±6.0. Last measured 7 Aug 2025stale: no new result in six months. Released 7 Aug 2025.
- Blended price
- $0.138/M
- $0.050 in · $0.400 out
- Output speed
- 176 tok/s
- First answer
- 96 s
- first token 2.19 s
- Context
- 400k
- Overall index51
- Reasoning57
- Agents & tools48
- Knowledge50
- Instruction following66
- Multimodal45
- Long context48
- Composite45
Reasoning-effort configurations
The same model behaves differently depending on how much it is allowed to think. Each row is one setting, scored only on the benchmarks that were run at that setting. “Default” means the publisher did not say which setting was used.
| Effort | Index | Rank | Speed | First answer | Chat reply cost | Categories |
|---|---|---|---|---|---|---|
| minimal | 36.4 | #578 | 168 tok/s | 0.88 s | $0.0001 | Agents & tools 40 · Composite 38 · Instruction following 36 · Knowledge 34 · Long context 35 · Maths 32 · Multimodal 15 · Reasoning 33 |
| low | 39.8 | #519 | 176 tok/s | 96 s | $0.0001 | Coding 34 · Maths 35 · Reasoning 36 |
| medium | 46.5 | #367 | 167 tok/s | 49 s | $0.0001 | Agents & tools 35 · Coding 35 · Composite 45 · Instruction following 65 · Knowledge 52 · Long context 37 · Maths 62 · Multimodal 42 · Reasoning 40 |
| high | 45.9 | #378 | 176 tok/s | 96 s | $0.0001 | Coding 48 · Human preference 51 · Knowledge 35 · Maths 48 · Multimodal 48 · Reasoning 42 |
| defaultbest | 51.1 | #270 | 176 tok/s | 96 s | $0.0001 | Agents & tools 48 · Composite 45 · Instruction following 66 · Knowledge 50 · Long context 48 · Multimodal 45 · Reasoning 57 |
Benchmark results
One column per reasoning effort. Rank is among every configuration of every model on that benchmark. Hover a score for the run it came from.
Reasoning
| Benchmark | minimal | low | medium | high | default | Source |
|---|---|---|---|---|---|---|
| GPQA Diamond | 48.5%#205 | 57.6%#181 | 67.4%#161 | 69.4%#158 | – | Epoch AI Benchmarking Hub |
| LMArena Hard Prompts | – | – | – | 1355#189 | – | LMArena |
| GPQA Diamond (AA)not in index | 42.8%#438 | – | 67.0%#311 | – | 67.6%#304 | Artificial Analysis |
| Humanity's Last Exam (AA)not in index | 4.0%#467 | – | 8.7%#277 | – | 9.5%#267 | Artificial Analysis |
| GPQA Diamond (Vals)not in index | – | – | – | 63.4%#102 | – | Vals AI |
| Kagi LLM Benchmark | – | – | – | – | 62.2%#46 | Kagi LLM Benchmark |
| ARC-AGI-1 | 1.5%#184 | 4.0%#181 | 20.7%#157 | 16.7%#160 | – | ARC Prize |
| ARC-AGI-2 | 0.0%#174 | 0.0%#174 | 0.9%#162 | 2.6%#137 | – | ARC Prize |
Coding
| Benchmark | minimal | low | medium | high | default | Source |
|---|---|---|---|---|---|---|
| WeirdML | – | 25.9%#128 | – | 38.1%#111 | – | WeirdML |
| LMArena Coding | – | – | – | 1384#188 | – | LMArena |
| LiveCodeBench | – | – | – | 70.2%#86 | – | Vals AI |
| SWE-bench Verified (bash only) | – | – | 34.8%#36 | – | – | SWE-bench |
| SWE-bench Verified (any scaffold)not in index | – | – | 34.8%#54 | – | – | SWE-bench |
Agents & tools
| Benchmark | minimal | low | medium | high | default | Source |
|---|---|---|---|---|---|---|
| Terminal-Bench | – | – | 11.5%#63 | – | 21.8%#53 | Terminal-Bench |
| Terminal-Bench Hard | 6.8%#240 | – | 17.4%#173 | – | 12.1%#202 | Artificial Analysis |
| τ²-Bench Telecom (AA)not in index | 25.7%#279 | – | 30.4%#251 | – | 36.5%#223 | Artificial Analysis |
| BFCL Overall | – | – | – | – | 51.5%#22 | Berkeley Function Calling Leaderboard |
Maths
| Benchmark | minimal | low | medium | high | default | Source |
|---|---|---|---|---|---|---|
| FrontierMath Tiers 1–3 | 1.8%#99 | 6.0%#94 | – | 20.0%#74 | – | Epoch AI Benchmarking Hub |
| FrontierMath Tier 4 | – | – | – | 2.4%#53 | – | Epoch AI Benchmarking Hub |
| OTIS Mock AIME | 35.6%#185 | 46.7%#169 | 74.2%#113 | 81.1%#98 | – | Epoch AI Benchmarking Hub |
| MATH Level 5 | – | – | 95.2%#13 | 94.9%#15 | – | Epoch AI Benchmarking Hub |
| ProofBench | – | – | – | 12.0%#48 | – | Vals AI |
| AIME (Vals) | – | – | – | 81.2%#45 | – | Vals AI |
| MGSM | – | – | – | 89.3%#45 | – | Vals AI |
Knowledge
| Benchmark | minimal | low | medium | high | default | Source |
|---|---|---|---|---|---|---|
| SimpleQA Verified | – | – | – | 11.7%#71 | – | Epoch AI Benchmarking Hub |
| AA-Omniscience | -64.1#415 | – | -25.8#200 | – | -28.7#210 | Artificial Analysis |
| MMLU-Pro | – | – | – | 76.1%#103 | – | Vals AI |
| LegalBench | – | – | – | 50.1%#134 | – | Vals AI |
| TaxEval | – | – | – | 67.4%#105 | – | Vals AI |
| MedQA | – | – | – | 93.3%#23 | – | Vals AI |
Instruction following
| Benchmark | minimal | low | medium | high | default | Source |
|---|---|---|---|---|---|---|
| IFBench | 32.5%#326 | – | 65.9%#100 | – | 67.5%#84 | Artificial Analysis |
Human preference
| Benchmark | minimal | low | medium | high | default | Source |
|---|---|---|---|---|---|---|
| LMArena Text | – | – | – | 1337#192 | – | LMArena |
Multimodal
| Benchmark | minimal | low | medium | high | default | Source |
|---|---|---|---|---|---|---|
| LMArena Vision | – | – | – | 1159#94 | – | LMArena |
| MMMU-Pro | 31.8%#237 | – | 58.2%#185 | – | 61.0%#174 | Artificial Analysis |
Long context
| Benchmark | minimal | low | medium | high | default | Source |
|---|---|---|---|---|---|---|
| Fiction.LiveBench 120k | – | – | 21.9%#33 | – | – | Fiction.live |
| AA-LCR | 20.0%#364 | – | 43.7%#274 | – | 45.0%#269 | Artificial Analysis |
Composite
| Benchmark | minimal | low | medium | high | default | Source |
|---|---|---|---|---|---|---|
| Epoch Capabilities Indexnot in index | – | – | – | – | 139.4#102 | Epoch AI Benchmarking Hub |
| AA Intelligence Index | 7.1#414 | – | 12.5#261 | – | 13.0#254 | Artificial Analysis |
Where to run it
Every provider serving this model through OpenRouter, with throughput and first-token latency measured on live traffic over the last 30 minutes and each provider’s own price. Purple marks the best in each column.
| Provider | Speed | First token | Input $/M | Output $/M | Context | Quantisation |
|---|---|---|---|---|---|---|
| OpenAI Flex | 93 tok/s | 1.29 s | $0.025 | $0.200 | 400k | – |
| OpenAI | 90 tok/s | 2.48 s | $0.050 | $0.400 | 400k | – |
| Azure (EU) | 83 tok/s | 2.76 s | $0.055 | $0.440 | 400k | – |
| Azure | 55 tok/s | 1.91 s | $0.050 | $0.400 | 400k | – |
Price history
Listed price per 1M tokens over time, as recorded by OpenRouter for the provider with the longest history.
What a task costs
Estimates from list price, output speed and time to first answer for the best configuration. “With caching” assumes three-quarters of the input is served from the prompt cache at $0.005 per 1M. Reasoning tokens are not modelled.
| Workload | Tokens in / out | Cost | With caching | Time |
|---|---|---|---|---|
| Chat reply | 400 / 300 | $0.0001 | $0.0001 | 1.6 min |
| Summarise a 30-page report | 12,000 / 600 | $0.0008 | $0.0004 | 1.7 min |
| Code edit | 6,000 / 1,500 | $0.0009 | $0.0007 | 1.7 min |
| Agentic coding session | 60,000 / 4,000 | $0.0046 | $0.0026 | 2.0 min |
| Structured extraction | 2,000 / 200 | $0.0002 | $0.0001 | 1.6 min |
See also
Data as of 9 Sept 2026. Compare these configurations.