OpenAIGPT-5.1
GPT-5.1 is an OpenAI proprietary reasoning model, released 12 Nov 2025. Its best configuration (high reasoning effort) ranks #170 of 785 on the BenchLeader Index at 57.7 ±3.5, in the upper half. The ± is the point: 238 other configurations score within that range, so they and this one cannot be told apart on quality alone — price and speed are what separate them. It scores highest in instruction following (71) and lowest in agents & tools (42). At $3.44 per million tokens blended it is among the most expensive ranked models. Output speed of 112 tokens per second puts it faster than most, with a first answer in 33.1 s. It has been measured at 6 reasoning-effort settings; this summary describes the best-scoring one, and the tabs above switch between them. Last measured 22 Oct 2026.
- Blended price
- $3.44/M
- $1.25 in · $10.00 out
- Output speed
- 112 tok/s
- measured by Artificial Analysis
- First answer
- 33 s
- first token 1.18 s
- Context
- 400k
- Full answer
- 38 s
- median, reasoning included
- Cost per run
- –
- Released
- 12 Nov 2025
- Overall index58
- Reasoning53
- Coding56
- Agents & tools42
- Maths59
- Knowledge60
- Instruction following71
- Human preference64
- Multimodal58
- Long context65
- Composite57
Versions
OpenAI has shipped 13 models under this name. Each is ranked on its own results; a newer version often has fewer results so far, which holds its index nearer the average until more arrive.
| Model | Released | Index | Rank |
|---|---|---|---|
| GPT-5.6 | 9 Jul 2026 | – | – |
| GPT-5.5xhigh | 23 Apr 2026 | 66.1 | #35 |
| GPT-5.4xhigh | 5 Mar 2026 | 64.7 | #45 |
| GPT-5.2xhigh | 11 Dec 2025 | 60.7 | #104 |
| GPT-5.1highthis page | 12 Nov 2025 | 57.7 | #170 |
| GPT-5high | 7 Aug 2025 | 57.9 | #166 |
| GPT-4.1 | 14 Apr 2025 | 48.3 | #428 |
| GPT-4.5 | 27 Feb 2025 | 49.7 | #388 |
| GPT 4 0125 | 25 Jan 2024 | – | – |
| GPT-4 1106 | 6 Nov 2023 | – | – |
| GPT 4 0613 | 13 Jun 2023 | 39.7 | #677 |
| GPT-4 | 14 Mar 2023 | 42.4 | #596 |
| GPT-3.5 | – | – | – |
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. “Not stated” collects results from publishers that did not say which setting they used; for a reasoning model that is usually its thinking mode, but we do not assume it. Pick a setting here or at the top of the page to see its price, speed and category scores.
| Effort | Index | Rank | Speed | First answer | Chat reply cost | Categories |
|---|---|---|---|---|---|---|
| no reasoning | 45.9 | #504 | 95 tok/s | 1.02 s | $0.0035 | Agents & tools 54 · Composite 44 · Instruction following 46 · Knowledge 46 · Long context 47 · Maths 40 · Multimodal 45 · Reasoning 38 |
| low | – | – | 52 tok/s | 1.18 s | $0.0035 | Maths 51 · Reasoning 34 |
| medium | 52.3 | #322 | 52 tok/s | 1.18 s | $0.0035 | Agents & tools 54 · Coding 51 · Maths 61 · Reasoning 48 |
| highbest | 57.7 | #170 | 112 tok/s | 33 s | $0.0035 | Agents & tools 42 · Coding 56 · Composite 57 · Human preference 64 · Instruction following 71 · Knowledge 60 · Long context 65 · Maths 59 · Multimodal 58 · Reasoning 53 |
| thinking | 54.1 | #261 | 52 tok/s | 1.18 s | $0.0035 | Instruction following 57 · Knowledge 59 · Multimodal 53 |
| not stated | 56.8 | #195 | 52 tok/s | 1.18 s | $0.0035 | Agents & tools 54 · Coding 45 · Human preference 62 · Instruction following 70 · Maths 59 · Multimodal 62 · Reasoning 56 |
How its index has moved
8 Sept 2026 to 23 Oct 2026- high
- not stated
- thinking
- medium
The index is recomputed from scratch every day, so a line moves when a new benchmark result lands, when a publisher revises a score, or when the models it is normalised against change. Early movement usually means the score is still settling.
Benchmark results
See all LLM benchmarks. 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.
Metadata
| Benchmark | high | Source | Trend |
|---|---|---|---|
| AA Openness Indexnot in index | 5.6#49 | Artificial Analysis |
Reasoning
Coding
| Benchmark | medium | high | not stated | Source | Trend |
|---|---|---|---|---|---|
| SWE-bench Verified (Epoch) | – | 68.0%#26 | – | Epoch AI Benchmarking Hub | |
| SciCodeeffort not stated | – | – | 43.3%#145 | SciCode | |
| WeirdML | – | 60.8%#47 | – | WeirdML | |
| GSO-Bench | – | 13.7%#20 | 13.7%#19 | GSO-Bench | |
| Arena Coding | – | 1492#91 | 1476#110 | Arena (LMArena) | |
| Arena WebDev | 1396#95 | – | 1342#111 | Arena (LMArena) | |
| LiveCodeBench | – | 86.5%#21 | – | Vals AI | |
| SWE-bench (Vals)not in index | – | 69.8%#61 | – | Vals AI | |
| SWE-bench Verified (bash only) | 66.0%#17 | – | – | SWE-bench | |
| SWE-bench Verified (any scaffold)not in index | 66.0%#19 | – | – | SWE-bench | |
| Vibe Code Bench v1.1not in index | – | 24.6%#78 | – | Vals AI | |
| ALE-Benchnot in index | – | 1192.2#42 | – | Epoch AI Benchmarking Hub | |
| Terminal-Bench 2.1 (AA)not in index | – | 52.4%#113 | – | Artificial Analysis |
Agents & tools
| Benchmark | no reasoning | low | medium | high | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|
| Terminal-Bench | – | – | 47.6%#37 | – | 47.6%#38 | Terminal-Bench | |
| Terminal-Bench Hard | 22.7%#153 | – | – | 45.5%#34 | – | Artificial Analysis | |
| τ²-Bench Telecom (AA)not in index | 46.5%#205 | – | – | 81.9%#115 | – | Artificial Analysis | |
| MCP Atlas | – | – | – | 50.1%#31 | – | Scale AI SEAL | |
| τ³-Banking (AA) | – | – | – | 15.9%#113 | – | Artificial Analysis | |
| Poker Agentnot in indexeffort not stated | – | – | – | – | 1038.6%#8 | Vals AI | |
| Terminal-Bench 2.0 (Vals)not in index | – | – | – | 44.9%#24 | – | Vals AI | |
| DeepResearch Benchnot in index | – | 42.8%#33 | – | – | – | DeepResearch Bench | |
| Vending-Bench 2not in indexeffort not stated | – | – | – | – | 1473.4#52 | Andon Labs |
Maths
| Benchmark | no reasoning | low | medium | high | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|
| OTIS Mock AIME | 37.8%#200 | 63.9%#156 | 85.6%#96 | 88.6%#81 | – | Epoch AI Benchmarking Hub | |
| AIME (Vals) | – | – | – | 93.3%#15 | – | Vals AI | |
| MGSM | – | – | – | 93.0%#12 | – | Vals AI | |
| MathArena Apexpossible contamination | – | – | – | 1.0%#35 | – | MathArena | |
| Omni-MATH (HELM)effort not stated | – | – | – | – | 46.4%#23 | HELM Capabilities | |
| Arena Maths | – | – | – | 1458#66 | 1423#117 | Arena (LMArena) |
Knowledge
| Benchmark | no reasoning | high | thinking | not stated | Source | Trend |
|---|---|---|---|---|---|---|
| SimpleQA Verified | – | 48.0%#33 | – | – | Epoch AI Benchmarking Hub | |
| AA-Omniscience | -34.5#299 | 5.4#105 | – | – | Artificial Analysis | |
| MMLU-Pro | – | 86.4%#41 | – | – | Vals AI | |
| LegalBench | – | 85.7%#20 | – | – | Vals AI | |
| CorpFin | – | 63.8%#40 | – | – | Vals AI | |
| TaxEval | – | 74.9%#25 | – | – | Vals AI | |
| MedQA | – | 96.4%#2 | – | – | Vals AI | |
| PRBench Finance | – | – | 48.0%#14 | – | Scale AI SEAL | |
| PRBench Legal | – | – | 49.3%#13 | – | Scale AI SEAL | |
| MultiNRCpossible contamination | – | – | 49.0%#10 | – | Scale AI SEAL | |
| MMLU-Pro (HELM)not in indexeffort not stated | – | – | – | 57.9%#51 | HELM Capabilities | |
| CaseLaw v2not in index | – | 73.4%#2 | – | – | Vals AI | |
| MedCodenot in index | – | 52.7%#13 | – | – | Vals AI | |
| MedScribenot in index | – | 88.1%#14 | – | – | Vals AI | |
| AA-Omniscience: accuracynot in index | 29.5%#188 | 37.7%#140 | – | – | Artificial Analysis | |
| AA-Omniscience: non-hallucinationnot in index | 9.4%#430 | 48.1%#129 | – | – | Artificial Analysis | |
| Global-MMLU-Lite (AA)not in index | – | 90.6%#7 | – | – | Artificial Analysis |
Instruction following
| Benchmark | no reasoning | high | thinking | not stated | Source | Trend |
|---|---|---|---|---|---|---|
| IFBench | 43.2%#227 | 72.9%#45 | – | – | Artificial Analysis | |
| MultiChallenge | – | – | 63.4%#10 | – | Scale AI SEAL | |
| TutorBench | – | – | 54.1%#7 | – | Scale AI SEAL | |
| IFEval (HELM)effort not stated | – | – | – | 93.5%#3 | HELM Capabilities | |
| WildBench (HELM)effort not stated | – | – | – | 86.3%#2 | HELM Capabilities | |
| Arena Instruction Followingnot in index | – | 1450#64 | – | 1427#99 | Arena (LMArena) |
Human preference
| Benchmark | high | not stated | Source | Trend |
|---|---|---|---|---|
| Arena Text | 1456#69 | 1439#96 | Arena (LMArena) | |
| Arena Creative Writingnot in index | 1429#74 | 1408#93 | Arena (LMArena) | |
| Arena Multi-turnnot in index | 1464#73 | 1448#88 | Arena (LMArena) | |
| Arena Longer Queriesnot in index | 1461#73 | 1443#101 | Arena (LMArena) |
Multimodal
| Benchmark | no reasoning | medium | high | thinking | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|
| Arena Vision | – | – | 1250#64 | – | 1235#75 | Arena (LMArena) | |
| MMMU-Pro | 62.4%#195 | – | 75.5%#98 | – | – | Artificial Analysis | |
| VISTA | – | – | – | 43.8%#31 | – | Scale AI SEAL | |
| MMMU (validation)effort not stated | – | – | – | – | 85.4%#1 | MMMU | |
| MMMU-Pro (official)not in index | – | – | – | 79.0%#4 | 76.0%#7 | MMMU | |
| MMMU-Pro (Vals)not in index | – | – | 83.2%#33 | – | – | Vals AI | |
| MortgageTaxnot in index | – | – | 61.4%#58 | – | – | Vals AI | |
| SAGEnot in index | – | – | 43.2%#51 | – | – | Vals AI | |
| VPCTnot in index | – | 53.3%#9 | 58.7%#7 | – | – | Epoch AI Benchmarking Hub |
Long context
| Benchmark | no reasoning | high | not stated | Source | Trend |
|---|---|---|---|---|---|
| AA-LCR | 45.0%#338 | 80.0%#78 | – | Artificial Analysis | |
| Arena Documentnot in indexeffort not stated | – | – | 1416#43 | Arena (LMArena) |
Composite
| Benchmark | no reasoning | high | not stated | Source | Trend |
|---|---|---|---|---|---|
| Epoch Capabilities Indexnot in indexeffort not stated | – | – | 149.8#62 | Epoch AI Benchmarking Hub | |
| AA Intelligence Index v4.3.2 | 13.3#311 | 24.7#170 | – | Artificial Analysis | |
| HELM Capabilities meannot in indexeffort not stated | – | – | 65.6%#33 | HELM Capabilities |
Safety & honesty
| Benchmark | thinking | Source | Trend |
|---|---|---|---|
| FORTRESSnot in index | 25.7%#31 | Scale AI SEAL | |
| MASKnot in indexpossible contamination | 86.3%#16 | Scale AI SEAL |
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 | 159 tok/s | 0.64 s | $0.625 | $5.00 | 400k | – |
| Azure | 84 tok/s | 2.01 s | $1.25 | $10.00 | 400k | – |
| OpenAI | 60 tok/s | 1.15 s | $1.25 | $10.00 | 400k | – |
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 | 35.7 s |
| Summarize a 30-page report | 12,000 / 600 | $0.021 | $0.011 | 38.4 s |
| Code edit | 6,000 / 1,500 | $0.022 | $0.017 | 46.4 s |
| Agentic coding session | 60,000 / 4,000 | $0.115 | $0.064 | 1.1 min |
| Structured extraction | 2,000 / 200 | $0.0045 | $0.0028 | 34.8 s |
More from OpenAI
See also
Questions people ask
- What are GPT-5.1's benchmark scores?
- As of 25 Oct 2026, GPT-5.1 (high effort) scores 57.7 ±3.5 on the BenchLeader Index, strongest in instruction following (71) and weakest in agents & tools (42). That rests on 32 independent benchmark results, and every score on this page links to its publisher.
- How does GPT-5.1 rank?
- GPT-5.1 (high effort) is #170 of 785 ranked configurations on 25 Oct 2026. 238 other configurations score within its margin, so the exact position matters less than the band. The leader is Claude Opus 5.5 (max effort) at 72.6.
- How much does GPT-5.1 cost per 1M tokens?
- As of 25 Oct 2026, GPT-5.1 costs $1.25 per 1M input tokens and $10.00 per 1M output tokens, $3.44 blended (three input tokens per output token).
- When was GPT-5.1 released?
- OpenAI released GPT-5.1 on 12 Nov 2025. BenchLeader has results for it at 6 reasoning-effort settings.
Data as of 25 Oct 2026. Compare these configurations.
Cite as: BenchLeader, “GPT-5.1: benchmarks, pricing, speed and rank”, data as of 25 Oct 2026, https://www.benchleader.com/models/gpt-5-1.