OpenAIReasoning modelGPT-5
Best configuration ranks #100 of 610 on the BenchLeader Index at 59.5 ±3.5. Last measured 2 Sept 2026. Released 7 Aug 2025.
- Blended price
- $3.44/M
- $1.25 in · $10.00 out
- Output speed
- 76 tok/s
- First answer
- 74 s
- first token 6.54 s
- Context
- 400k
- Overall index60
- Reasoning61
- Coding61
- Agents & tools51
- Knowledge62
- Instruction following68
- Human preference62
- Multimodal60
- Long context66
- Composite58
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 | 46.0 | #374 | 79 tok/s | 1.35 s | $0.0035 | Agents & tools 49 · Composite 43 · Instruction following 47 · Knowledge 48 · Maths 40 · Multimodal 46 · Reasoning 40 |
| low | 53.1 | #233 | 72 tok/s | 9.74 s | $0.0035 | Agents & tools 57 · Composite 55 · Instruction following 65 · Knowledge 59 · Maths 48 · Multimodal 58 · Reasoning 37 |
| medium | 58.5 | #122 | 81 tok/s | 42 s | $0.0035 | Agents & tools 57 · Coding 52 · Composite 58 · Instruction following 69 · Knowledge 59 · Long context 71 · Maths 66 · Multimodal 59 · Reasoning 49 |
| high | 54.5 | #195 | 76 tok/s | 74 s | $0.0035 | Agents & tools 43 · Coding 55 · Human preference 62 · Knowledge 59 · Maths 57 · Multimodal 54 · Reasoning 55 |
| thinking | – | – | 76 tok/s | 74 s | $0.0035 | Instruction following 60 |
| defaultbest | 59.5 | #100 | 76 tok/s | 74 s | $0.0035 | Agents & tools 51 · Coding 61 · Composite 58 · Human preference 62 · Instruction following 68 · Knowledge 62 · Long context 66 · Multimodal 60 · Reasoning 61 |
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 | thinking | default | Source | Trend |
|---|---|---|---|---|---|---|---|---|
| GPQA Diamond | 71.7%#150 | – | 85.3%#76 | 86.2%#70 | – | – | Epoch AI Benchmarking Hub | |
| Humanity's Last Exam | – | – | – | 25.3%#11 | – | 25.3%#11 | Scale AI / CAIS | |
| SimpleBench | – | – | – | 56.7%#36 | – | – | SimpleBench | |
| LMArena Hard Prompts | – | – | – | 1448#95 | – | 1449#93 | LMArena | |
| GPQA Diamond (AA)not in index | 67.3%#306 | 80.8%#184 | 84.2%#147 | – | – | 85.3%#127 | Artificial Analysis | |
| Humanity's Last Exam (AA)not in index | 6.0%#333 | 19.6%#163 | 25.4%#133 | – | – | 28.5%#115 | Artificial Analysis | |
| GPQA Diamond (Vals)not in index | – | – | – | 85.6%#44 | – | – | Vals AI | |
| Kagi LLM Benchmark | – | – | – | – | – | 72.7%#21 | Kagi LLM Benchmark | |
| ARC-AGI-1 | 6.0%#172 | 44.0%#128 | 56.2%#116 | 65.7%#101 | – | – | ARC Prize | |
| ARC-AGI-2 | 0.0%#174 | 1.9%#144 | 7.5%#111 | 9.9%#106 | – | – | ARC Prize |
Coding
| Benchmark | minimal | low | medium | high | thinking | default | Source | Trend |
|---|---|---|---|---|---|---|---|---|
| SWE-bench Verified (Epoch) | – | – | 71.5%#23 | 73.5%#19 | – | – | Epoch AI Benchmarking Hub | |
| SciCode | – | – | – | – | – | 42.9%#94 | SciCode | |
| WeirdML | – | – | – | 60.7%#45 | – | 39.8%#102 | WeirdML | |
| GSO-Bench | – | – | – | 6.9%#20 | – | – | GSO-Bench | |
| LMArena Coding | – | – | – | 1470#101 | – | 1463#110 | LMArena | |
| LMArena WebDev | – | – | 1420#64 | – | – | – | LMArena | |
| LiveCodeBench | – | – | – | 85.9%#27 | – | – | Vals AI | |
| IOI | – | – | – | 20.0%#24 | – | – | Vals AI | |
| SWE-bench (Vals)not in index | – | – | – | 69.0%#64 | – | – | Vals AI | |
| SWE-Bench Pro | – | – | – | 41.8%#10 | – | – | Scale AI SEAL | |
| Aider Polyglot | – | – | – | – | – | 88.0%#1 | Aider polyglot leaderboard | |
| SWE-bench Verified (bash only) | – | – | 65.0%#19 | – | – | – | SWE-bench | |
| SWE-bench Verified (any scaffold)not in index | – | – | – | – | – | 75.6%#7 | SWE-bench |
Agents & tools
| Benchmark | minimal | low | medium | high | thinking | default | Source | Trend |
|---|---|---|---|---|---|---|---|---|
| Terminal-Bench | – | – | 49.6%#26 | – | – | 49.6%#26 | Terminal-Bench | |
| GDPval | – | – | 34.8%#6 | – | – | – | OpenAI | |
| Remote Labor Index | – | – | – | – | – | 1.7%#10 | Scale AI / CAIS | |
| APEX-Agents | – | – | – | 18.3%#39 | – | 18.3%#39 | Mercor | |
| Terminal-Bench Hard | 18.2%#166 | 26.5%#126 | 37.9%#61 | – | – | 32.6%#95 | Artificial Analysis | |
| τ²-Bench Telecom (AA)not in index | 67.0%#160 | 84.2%#101 | 86.5%#85 | – | – | 84.8%#95 | Artificial Analysis |
Maths
| Benchmark | minimal | low | medium | high | thinking | default | Source | Trend |
|---|---|---|---|---|---|---|---|---|
| FrontierMath Tiers 1–3 | 18.3%#80 | 37.2%#49 | – | 55.4%#35 | – | – | Epoch AI Benchmarking Hub | |
| FrontierMath Tier 4 | – | – | – | 21.9%#38 | – | – | Epoch AI Benchmarking Hub | |
| OTIS Mock AIME | 46.7%#169 | – | 87.2%#69 | 91.4%#53 | – | – | Epoch AI Benchmarking Hub | |
| MATH Level 5 | – | – | 97.9%#2 | 98.1%#1 | – | – | Epoch AI Benchmarking Hub | |
| ProofBench | – | – | – | 18.0%#38 | – | – | Vals AI | |
| AIME (Vals) | – | – | – | 93.4%#14 | – | – | Vals AI | |
| MGSM | – | – | – | 92.8%#14 | – | – | Vals AI | |
| IMO 2025 | – | – | – | 38.1%#1 | – | – | MathArena | |
| MathArena Apex | – | – | – | 1.0%#35 | – | – | MathArena |
Knowledge
| Benchmark | minimal | low | medium | high | thinking | default | Source | Trend |
|---|---|---|---|---|---|---|---|---|
| SimpleQA Verified | – | – | – | 50.1%#21 | – | – | Epoch AI Benchmarking Hub | |
| AA-Omniscience | -33.8#229 | -10.8#146 | -10.9#147 | – | – | -8.7#130 | Artificial Analysis | |
| MMLU-Pro | – | – | – | 86.5%#40 | – | – | Vals AI | |
| LegalBench | – | – | – | 86.0%#14 | – | – | Vals AI | |
| CorpFin | – | – | – | 61.1%#59 | – | – | Vals AI | |
| TaxEval | – | – | – | 73.4%#48 | – | – | Vals AI | |
| MedQA | – | – | – | 96.3%#4 | – | – | Vals AI | |
| PRBench Finance | – | – | – | – | – | 51.3%#5 | Scale AI SEAL | |
| PRBench Legal | – | – | – | – | – | 49.0%#10 | Scale AI SEAL | |
| MultiNRC | – | – | – | – | – | 52.1%#9 | Scale AI SEAL |
Instruction following
| Benchmark | minimal | low | medium | high | thinking | default | Source | Trend |
|---|---|---|---|---|---|---|---|---|
| IFBench | 45.6%#199 | 66.6%#92 | 70.6%#66 | – | – | 73.1%#44 | Artificial Analysis | |
| MultiChallenge | – | – | – | – | 63.2%#7 | – | Scale AI SEAL | |
| TutorBench | – | – | – | – | – | 55.3%#4 | Scale AI SEAL |
Human preference
| Benchmark | minimal | low | medium | high | thinking | default | Source | Trend |
|---|---|---|---|---|---|---|---|---|
| LMArena Text | – | – | – | 1434#88 | – | 1427#97 | LMArena |
Multimodal
| Benchmark | minimal | low | medium | high | thinking | default | Source | Trend |
|---|---|---|---|---|---|---|---|---|
| LMArena Vision | – | – | – | 1209#70 | – | 1232#61 | LMArena | |
| MMMU-Pro | 62.1%#166 | 73.8%#93 | 74.3%#84 | – | – | 74.2%#86 | Artificial Analysis | |
| VISTA | – | – | – | – | – | 49.7%#11 | Scale AI SEAL |
Long context
| Benchmark | minimal | low | medium | high | thinking | default | Source | Trend |
|---|---|---|---|---|---|---|---|---|
| Fiction.LiveBench 120k | – | – | 96.9%#2 | – | – | – | Fiction.live | |
| AA-LCR | – | – | 76.0%#102 | – | – | 78.2%#81 | Artificial Analysis |
Composite
| Benchmark | minimal | low | medium | high | thinking | default | Source | Trend |
|---|---|---|---|---|---|---|---|---|
| Epoch Capabilities Indexnot in index | – | – | – | – | – | 150#45 | Epoch AI Benchmarking Hub | |
| AA Intelligence Index | 11.4#286 | 20.8#167 | 22.9#138 | – | – | 23.0#136 | Artificial Analysis |
Where it wins
Benchmarks where this configuration ranks in the top five of every configuration measured.
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 | 62 tok/s | 3.81 s | $1.25 | $10.00 | 400k | – |
| Azure | 48 tok/s | 9.27 s | $1.25 | $10.00 | 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 $0.125 per 1M. Reasoning tokens are not modelled.
| Workload | Tokens in / out | Cost | With caching | Time |
|---|---|---|---|---|
| Chat reply | 400 / 300 | $0.0035 | $0.0032 | 1.3 min |
| Summarise a 30-page report | 12,000 / 600 | $0.021 | $0.011 | 1.4 min |
| Code edit | 6,000 / 1,500 | $0.022 | $0.017 | 1.6 min |
| Agentic coding session | 60,000 / 4,000 | $0.115 | $0.064 | 2.1 min |
| Structured extraction | 2,000 / 200 | $0.0045 | $0.0028 | 1.3 min |
See also
Data as of 9 Sept 2026. Compare these configurations.