OpenAIGPT-5.2
GPT-5.2 is an OpenAI proprietary reasoning model, released 11 Dec 2025. Its best configuration (xhigh reasoning effort) ranks #104 of 785 on the BenchLeader Index at 60.7 ±3.5, in the upper half. The ± is the point: 201 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 (74) and lowest in agents & tools (55). At $4.81 per million tokens blended it is among the most expensive ranked models. Output speed of 75 tokens per second puts it slower than most, with a first answer in 80.6 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 13 Oct 2026.
- Blended price
- $4.81/M
- $1.75 in · $14.00 out
- Output speed
- 75 tok/s
- measured by Artificial Analysis
- First answer
- 81 s
- first token 2.21 s
- Context
- 400k
- Full answer
- 87 s
- median, reasoning included
- Cost per run
- –
- Released
- 11 Dec 2025
- Overall index61
- Reasoning61
- Coding63
- Agents & tools55
- Maths57
- Knowledge57
- Instruction following74
- Long context66
- Composite64
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-5.2xhighthis page | 11 Dec 2025 | 60.7 | #104 |
| GPT-5.1high | 12 Nov 2025 | 57.7 | #170 |
| GPT-5high | 7 Aug 2025 | 57.9 | #165 |
| GPT-4.1 | 14 Apr 2025 | 48.3 | #429 |
| 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 | 48.8 | #414 | 67 tok/s | 0.77 s | $0.0049 | Agents & tools 40 · Coding 50 · Composite 48 · Instruction following 49 · Knowledge 56 · Long context 47 · Maths 51 · Multimodal 49 · Reasoning 48 |
| low | 49.7 | #389 | 50 tok/s | 2.21 s | $0.0049 | Coding 50 · Knowledge 43 · Maths 58 · Reasoning 48 |
| medium | 57.6 | #173 | 50 tok/s | 2.21 s | $0.0049 | Agents & tools 61 · Coding 59 · Composite 59 · Instruction following 65 · Knowledge 52 · Long context 60 · Maths 65 · Multimodal 58 · Reasoning 54 |
| high | 56.0 | #214 | 50 tok/s | 2.21 s | $0.0049 | Agents & tools 59 · Coding 59 · Human preference 62 · Knowledge 44 · Maths 62 · Multimodal 58 · Reasoning 56 |
| xhighbest | 60.7 | #104 | 75 tok/s | 81 s | $0.0049 | Agents & tools 55 · Coding 63 · Composite 64 · Instruction following 74 · Knowledge 57 · Long context 66 · Maths 57 · Reasoning 61 |
| not stated | 58.0 | #162 | 50 tok/s | 2.21 s | $0.0049 | Agents & tools 61 · Coding 53 · Human preference 67 · Instruction following 56 · Knowledge 56 · Maths 63 · Multimodal 59 · Reasoning 60 |
How its index has moved
8 Sept 2026 to 21 Oct 2026- xhigh
- not stated
- medium
- high
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.
Reasoning
| Benchmark | no reasoning | low | medium | high | xhigh | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|---|
| GPQA Diamond | 73.2%#158 | 82.7%#111 | 87.9%#65 | 88.2%#63 | 91.4%#32 | – | Epoch AI Benchmarking Hub | |
| Humanity's Last Exameffort not stated | – | – | – | – | – | 27.8%#14 | Scale AI / CAIS | |
| SimpleBench | – | – | – | 45.8%#52 | – | 45.8%#52 | SimpleBench | |
| Arena Hard Prompts | – | – | – | 1460#96 | – | 1497#38 | Arena (LMArena) | |
| LiveBench 1.0 Reasoningnot in index | – | – | – | 83.2%#44 | – | – | LiveBench | |
| GPQA Diamond (AA)not in index | 71.2%#288 | – | 86.4%#111 | – | 90.3%#63 | – | Artificial Analysis | |
| Humanity's Last Exam (AA)not in index | 8.0%#353 | – | 26.7%#178 | – | 37.7%#102 | – | Artificial Analysis | |
| GPQA Diamond (Vals)not in index | – | – | – | – | 91.7%#18 | – | Vals AI | |
| Kagi LLM Benchmarkeffort not stated | – | – | – | – | – | 73.3%#18 | Kagi LLM Benchmark | |
| ARC-AGI-1 | – | 55.7%#172 | 72.7%#139 | 78.7%#123 | 86.2%#105 | 94.5%#39 | ARC Prize | |
| ARC-AGI-2 | – | 9.7%#161 | 26.7%#138 | 43.3%#114 | 52.9%#104 | 72.9%#53 | ARC Prize | |
| CritPt | 0.6%#254 | – | 7.9%#141 | – | 11.6%#111 | – | Artificial Analysis | |
| SciPredictnot in indexeffort not stated | – | – | – | – | – | 20.6%#6 | Scale AI SEAL | |
| Chess Puzzlesnot in index | 2.0%#140 | 26.0%#76 | 36.0%#49 | 45.0%#29 | 48.0%#19 | – | Epoch AI Benchmarking Hub | |
| Mystery Game Puzzlesnot in index | 14.0%#97 | 10.0%#112 | 22.0%#55 | 23.0%#51 | – | – | Epoch AI Benchmarking Hub | |
| CL-benchnot in index | – | – | – | 18.1%#13 | – | 18.2%#12 | Epoch AI Benchmarking Hub | |
| LMCAnot in index | – | – | – | – | 43.9%#97 | – | Epoch AI Benchmarking Hub | |
| DTBenchnot in index | – | – | – | – | 90.9%#76 | – | Epoch AI Benchmarking Hub | |
| ForecastBenchnot in indexeffort not stated | – | – | – | – | – | 64.2%#31 | ForecastBench |
Coding
| Benchmark | no reasoning | low | medium | high | xhigh | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|---|
| SWE-bench Verified (Epoch) | – | – | – | 73.8%#17 | – | – | Epoch AI Benchmarking Hub | |
| WeirdML | 49.6%#70 | 49.6%#70 | 63.4%#42 | – | 72.2%#30 | – | WeirdML | |
| GSO-Bench | – | – | – | 27.4%#13 | – | – | GSO-Bench | |
| Arena Coding | – | – | – | 1489#95 | – | 1515#45 | Arena (LMArena) | |
| Arena WebDeveffort not stated | – | – | – | – | – | 1417#84 | Arena (LMArena) | |
| LiveBench 1.0 Codingnot in index | – | – | – | 76.1%#47 | – | – | LiveBench | |
| LiveCodeBench | – | – | – | – | 85.4%#31 | – | Vals AI | |
| SWE-bench (Vals)not in index | – | – | – | – | 75.8%#41 | – | Vals AI | |
| SWE-Bench Proeffort not stated | – | – | – | – | – | 29.9%#16 | Scale AI SEAL | |
| SWE-bench Verified (bash only) | – | – | – | 72.8%#7 | – | 69.0%#14 | SWE-bench | |
| SWE-bench Verified (any scaffold)not in index | – | – | – | 72.8%#10 | – | – | SWE-bench | |
| Vibe Code Bench v1.1not in index | – | – | – | – | 53.5%#57 | – | Vals AI | |
| SWE-Bench Pro (private)not in indexeffort not stated | – | – | – | – | – | 29.9%#16 | Scale AI SEAL | |
| ALE-Benchnot in index | – | – | 1249.8#37 | 1293.5#35 | – | – | Epoch AI Benchmarking Hub | |
| AlgoTunenot in index | – | – | 2.0#1 | – | – | – | AlgoTune |
Agents & tools
| Benchmark | no reasoning | low | medium | high | xhigh | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|---|
| Terminal-Bench | – | – | 64.9%#10 | – | – | 64.9%#9 | Terminal-Bench | |
| GDPval | 49.7%#1 | – | – | – | – | – | OpenAI | |
| Remote Labor Index | – | – | 2.5%#11 | – | – | 2.1%#12 | Scale AI / CAIS | |
| LiveBench Agentic Codingnot in index | – | – | – | 50.3%#44 | – | – | LiveBench | |
| Terminal-Bench Hard | 31.8%#101 | – | 43.2%#41 | – | 47.0%#27 | – | Artificial Analysis | |
| τ²-Bench Telecom (AA)not in index | 46.5%#205 | – | 74.3%#141 | – | 84.8%#97 | – | Artificial Analysis | |
| MCP Atlas | – | – | – | – | 67.6%#23 | – | Scale AI SEAL | |
| BFCL Overalleffort not stated | – | – | – | – | – | 55.9%#14 | Berkeley Function Calling Leaderboard | |
| τ²-bench | 61.6%#8 | – | – | 84.8%#3 | – | – | τ²-bench | |
| τ³-bench Bankingnot in index | 12.6%#24 | – | – | 32.2%#13 | – | – | τ²-bench | |
| τ³-Banking (AA) | 11.1%#139 | – | – | – | – | – | Artificial Analysis | |
| Poker Agentnot in indexeffort not stated | – | – | – | – | – | 1131.8%#1 | Vals AI | |
| Terminal-Bench 2.0 (Vals)not in index | – | – | – | 51.7%#18 | – | – | Vals AI | |
| DeepResearch Benchnot in index | – | 41.1%#35 | – | – | – | – | DeepResearch Bench | |
| METR Time Horizonsnot in index | – | – | – | 75.3%#4 | – | – | METR | |
| Vending-Bench 2not in indexeffort not stated | – | – | – | – | – | 3591.3#47 | Andon Labs |
Maths
| Benchmark | no reasoning | low | medium | high | xhigh | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|---|
| FrontierMath Tiers 1–3 | – | – | – | – | 67.4%#32 | – | Epoch AI Benchmarking Hub | |
| FrontierMath Tier 4 | – | – | – | – | 31.7%#39 | – | Epoch AI Benchmarking Hub | |
| OTIS Mock AIME | 62.2%#159 | 78.9%#120 | 93.9%#57 | 96.1%#40 | 96.1%#42 | – | Epoch AI Benchmarking Hub | |
| ProofBench | – | – | – | – | 15.0%#59 | – | Vals AI | |
| LiveBench Mathematicsnot in index | – | – | – | 93.2%#24 | – | – | LiveBench | |
| AIME (Vals) | – | – | – | – | 96.9%#2 | – | Vals AI | |
| MGSM | – | – | – | – | 94.0%#6 | – | Vals AI | |
| AIME 2026 | – | – | – | 98.3%#4 | – | – | MathArena | |
| HMMT February 2026 | – | – | – | 97.0%#3 | – | – | MathArena | |
| MathArena Apexpossible contamination | – | – | – | 13.5%#18 | – | – | MathArena | |
| Arena Maths | – | – | – | 1456#67 | – | 1454#71 | Arena (LMArena) |
Knowledge
| Benchmark | no reasoning | low | medium | high | xhigh | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|---|
| SimpleQA Verified | – | 32.8%#61 | 32.7%#63 | 34.3%#53 | 37.1%#49 | – | Epoch AI Benchmarking Hub | |
| LiveBench 1.0 Data Analysisnot in index | – | – | – | 78.2%#27 | – | – | LiveBench | |
| AA-Omniscience | -12.3#210 | – | 0.3#133 | – | -0.9#144 | – | Artificial Analysis | |
| MMLU-Pro | – | – | – | – | 86.2%#44 | – | Vals AI | |
| LegalBench | – | – | – | – | 82.8%#64 | – | Vals AI | |
| CorpFin | – | – | – | – | 65.9%#25 | – | Vals AI | |
| TaxEval | – | – | – | – | 75.8%#11 | – | Vals AI | |
| MedQA | – | – | – | – | 94.1%#19 | – | Vals AI | |
| MultiNRCeffort not statedpossible contamination | – | – | – | – | – | 42.2%#17 | Scale AI SEAL | |
| CaseLaw v2not in index | – | – | – | – | 66.0%#8 | – | Vals AI | |
| MedCodenot in index | – | – | – | – | 49.8%#20 | – | Vals AI | |
| MedScribenot in index | – | – | – | – | 84.4%#35 | – | Vals AI | |
| DrugDiscoveryBenchnot in index | – | – | – | – | 29.3%#16 | – | Scale AI SEAL | |
| EBR-benchnot in index | – | – | – | – | 23.0%#16 | – | Epoch AI Benchmarking Hub | |
| AA-Omniscience: accuracynot in index | 30.9%#177 | – | 38.3%#136 | – | 44.3%#94 | – | Artificial Analysis | |
| AA-Omniscience: non-hallucinationnot in index | 37.7%#175 | – | 38.4%#170 | – | 18.8%#292 | – | Artificial Analysis |
Instruction following
| Benchmark | no reasoning | medium | high | xhigh | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|
| LiveBench Languagenot in index | – | – | 79.8%#32 | – | – | LiveBench | |
| LiveBench Instruction Followingnot in index | – | – | 61.8%#55 | – | – | LiveBench | |
| IFBench | 47.4%#188 | 65.2%#103 | – | 75.4%#31 | – | Artificial Analysis | |
| TutorBencheffort not stated | – | – | – | – | 53.5%#11 | Scale AI SEAL | |
| Arena Instruction Followingnot in index | – | – | 1427#98 | – | 1457#55 | Arena (LMArena) |
Human preference
| Benchmark | high | not stated | Source | Trend |
|---|---|---|---|---|
| Arena Text | 1437#98 | 1476#36 | Arena (LMArena) | |
| Arena Creative Writingnot in index | 1386#127 | 1433#68 | Arena (LMArena) | |
| Arena Multi-turnnot in index | 1442#99 | 1495#16 | Arena (LMArena) | |
| Arena Longer Queriesnot in index | 1434#115 | 1472#58 | Arena (LMArena) |
Multimodal
| Benchmark | no reasoning | medium | high | xhigh | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|
| Arena Vision | – | – | 1246#67 | – | 1268#47 | Arena (LMArena) | |
| MMMU-Pro | 65.8%#172 | 74.6%#110 | – | – | – | Artificial Analysis | |
| VISTAeffort not stated | – | – | – | – | 46.6%#20 | Scale AI SEAL | |
| MMMU-Pro (Vals)not in index | – | – | – | 86.7%#17 | – | Vals AI | |
| MortgageTaxnot in index | – | – | – | 67.1%#29 | – | Vals AI | |
| SAGEnot in index | – | – | – | 49.3%#22 | – | Vals AI | |
| VPCTnot in index | – | – | 67.0%#4 | 84.0%#2 | – | Epoch AI Benchmarking Hub |
Long context
| Benchmark | no reasoning | medium | high | xhigh | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|
| AA-LCR | 45.7%#335 | 70.3%#211 | – | 82.7%#36 | – | Artificial Analysis | |
| Arena Documentnot in index | – | – | 1423#41 | – | 1425#40 | Arena (LMArena) |
Composite
| Benchmark | no reasoning | medium | high | xhigh | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|
| Epoch Capabilities Indexnot in indexeffort not stated | – | – | – | – | 153.7#45 | Epoch AI Benchmarking Hub | |
| LiveBench 1.0not in index | – | – | 74.6%#40 | – | – | LiveBench | |
| AA Intelligence Index v4.3.2 | 17.0#264 | 26.5#151 | – | 30.4#119 | – | Artificial Analysis | |
| Vals RSI Indexnot in index | – | – | – | 15.8%#22 | – | Vals AI |
Safety & honesty
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| FORTRESSnot in index | 17.5%#45 | Scale AI SEAL | |
| MASKnot in indexpossible contamination | 86.7%#13 | Scale AI SEAL | |
| PropensityBenchnot in index | 34.4%#9 | 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 |
|---|---|---|---|---|---|---|
| Azure | 59 tok/s | 1.42 s | $1.75 | $14.00 | 400k | – |
| OpenAI | 57 tok/s | 2.93 s | $1.75 | $14.00 | 400k | – |
| OpenAI | 47 tok/s | 1.07 s | $3.50 | $28.00 | 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 $0.175 per 1M. Reasoning tokens are not modelled.
| Workload | Tokens in / out | Cost | With caching | Time |
|---|---|---|---|---|
| Chat reply | 400 / 300 | $0.0049 | $0.0044 | 1.4 min |
| Summarize a 30-page report | 12,000 / 600 | $0.029 | $0.015 | 1.5 min |
| Code edit | 6,000 / 1,500 | $0.032 | $0.024 | 1.7 min |
| Agentic coding session | 60,000 / 4,000 | $0.161 | $0.090 | 2.2 min |
| Structured extraction | 2,000 / 200 | $0.0063 | $0.0039 | 1.4 min |
More from OpenAI
Compare GPT-5.2 with…
See also
Questions people ask
- What are GPT-5.2's benchmark scores?
- As of 24 Oct 2026, GPT-5.2 (xhigh effort) scores 60.7 ±3.5 on the BenchLeader Index, strongest in instruction following (74) and weakest in agents & tools (55). That rests on 24 independent benchmark results, and every score on this page links to its publisher.
- How does GPT-5.2 rank?
- GPT-5.2 (xhigh effort) is #104 of 785 ranked configurations on 24 Oct 2026. 201 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-5.2 cost per 1M tokens?
- As of 24 Oct 2026, GPT-5.2 costs $1.75 per 1M input tokens and $14.00 per 1M output tokens, $4.81 blended (three input tokens per output token).
- Is GPT-5.2 better than GPT-5.6 Sol?
- As of 24 Oct 2026, GPT-5.6 Sol (max) leads on quality (68.0 vs 60.7 on the BenchLeader Index), but GPT-5.2 (xhigh) costs 1.7× less, so it is the better buy unless you need the extra quality. The head-to-head page has the benchmark-by-benchmark detail.
- When was GPT-5.2 released?
- OpenAI released GPT-5.2 on 11 Dec 2025. BenchLeader has results for it at 6 reasoning-effort settings.
Data as of 24 Oct 2026. Compare these configurations.
Cite as: BenchLeader, “GPT-5.2: benchmarks, pricing, speed and rank”, data as of 24 Oct 2026, https://www.benchleader.com/models/gpt-5-2.