OpenAI
availableShows if the model has enough results for an index.GPT-5.2
GPT-5.2 is a reasoning model from OpenAI. 44 benchmarks count toward its score, in 8 categories.
IndexOverall score. 50 is the middle.62.2 ±2.4
CoverageShare of the index weight with results.100%
SpeedOutput tokens per second.43/s
Input / 1MUS dollars per 1M input tokens.$1.75 batch $0.875
Output / 1MUS dollars per 1M output tokens.$14 batch $7 US dollars per 1M output tokens in a batch.
ContextMaximum tokens in one request.400K
EloLMArena rating and rank.1412 (#125)
50 is the middle of the board. The range shows the doubt in the index. Batch work costs less.
78,967 votes. Elo shows what people prefer. It does not change the score.
CapabilitiesScore per category. 50 is the middle.
50 is the middleResults
44 counted| BenchmarkThe test name. | CategoryThe capability that the test measures. | ResultThe score from the publisher. | IndexThis result on the index scale. | RunThe settings of the run. | DateDate of the result. | Published byThe source of the result. |
|---|---|---|---|---|---|---|
| Artificial Analysis AIME 2025 | Math | 99.0% | — | — | — | Artificial Analysis |
| AIME | Math | 96.9% | 58.9 | — | 16 Apr 2026 | Vals AI |
| OTIS Mock AIME 2024-2025 | Math | 96.1% | 65.3 | xhigh effort | — | Epoch AI |
| MGSM | Multilingual | 94.0% | — | — | 9 Jan 2026 | Vals AI |
| LiveBench Mathematics | Math | 93.2% | 72.4 | high effort | 25 Jun 2026 | LiveBench |
| Graduate-Level Google-Proof Q&A | Knowledge | 92.4% | 63.5 | — | — | David Rein et al. |
| GPQA Diamond | Knowledge | 91.7% | 62.9 | — | 1 Sept 2026 | Vals AI |
| GPQA diamond | Knowledge | 91.4% | 62.6 | xhigh effort | — | Epoch AI |
| Artificial Analysis GPQA Diamond | Knowledge | 90.3% | 61.1 | — | — | Artificial Analysis |
| τ²-bench Telecom | Agentic | 89.7% | 63.2 | high effort · Sierra | 2 Mar 2026 | Sierra Research |
| MMMU Pro | Multimodal | 86.7% | 68.4 | — | 1 Sept 2026 | Vals AI |
| MMLU Pro | Knowledge | 86.2% | 56.4 | — | 1 Sept 2026 | Vals AI |
| ARC-AGI-1 (semi-private) | Reasoning | 86.2% | 67.9 | xhigh effort | — | ARC Prize Foundation |
| LiveCodeBench | Coding | 85.4% | 61.9 | — | 1 Sept 2026 | Vals AI |
| τ²-Bench Tool-Agent-User Evaluation | Agentic | 84.8% | 59.7 | — | — | Victor Barres et al. |
| LiveBench Reasoning | Reasoning | 83.2% | 71.6 | high effort | 25 Jun 2026 | LiveBench |
| MathVision | Multimodal | 83.0% | — | — | — | Qwen |
| τ²-bench Airline | Agentic | 83.0% | 58.4 | high effort · Sierra | 2 Mar 2026 | Sierra Research |
| Artificial Analysis Long Context Reasoning | Reasoning | 82.7% | 65.5 | — | — | Artificial Analysis |
| CharXiv Reasoning | Multimodal | 82.1% | 59.6 | — | — | CharXiv authors |
| τ²-bench Retail | Agentic | 81.6% | 57.4 | high effort · Sierra | 30 Apr 2026 | Sierra Research |
| Software Engineering Benchmark Verified | Coding | 80.0% | 61.7 | — | — | Carlos E. Jimenez et al. |
| LiveBench Language | Knowledge | 79.8% | 66.8 | high effort | 25 Jun 2026 | LiveBench |
| Massive Multi-discipline Multimodal Understanding Pro | Multimodal | 79.5% | 56.7 | — | — | MMMU-Pro authors |
| LiveBench Data Analysis | Reasoning | 78.2% | 64.5 | high effort | 25 Jun 2026 | LiveBench |
| LiveBench Coding | Coding | 76.1% | 64.0 | high effort | 25 Jun 2026 | LiveBench |
| V* | Multimodal | 75.9% | 31.9 | — | — | Z.AI |
| SWE-bench | Coding | 75.8% | 58.3 | — | 1 Sept 2026 | Vals AI |
| Artificial Analysis IFBench | Instruction | 75.4% | 65.6 | — | — | Artificial Analysis |
| SWE-Bench verified | Coding | 73.8% | 56.7 | high effort | — | Epoch AI |
| SWE-bench Verified | Coding | 72.3% | 55.5 | high effort · mini-SWE-agent | 1 Sept 2026 | SWE-bench team |
| FrontierMath-Tiers-1-3-v2-Private | Math | 67.4% | 69.4 | xhigh effort | — | Epoch AI |
| SWE-bench Multilingual | Multilingual | 66.7% | — | high effort · mini-SWE-agent | 20 Feb 2026 | SWE-bench team |
| SWE-bench Multilingual | Coding | 66.7% | — | high effort · mini-SWE-agent | 2 Sept 2026 | SWE-bench team |
| BrowseComp | Agentic | 65.8% | 55.1 | — | — | OpenAI |
| Terminal-Bench 1.0 | Agentic | 63.8% | 65.6 | — | 12 Jan 2026 | Vals AI |
| EuroEval Swedish | Multilingual | 63.5% | 81.6 | — | — | EuroEval |
| EuroEval Portuguese | Multilingual | 62.5% | 80.4 | — | — | EuroEval |
| EuroEval French | Multilingual | 62.0% | 79.7 | — | — | EuroEval |
| LiveBench Instruction Following | Instruction | 61.8% | 57.6 | high effort | 25 Jun 2026 | LiveBench |
| EuroEval Italian | Multilingual | 60.1% | 77.3 | — | — | EuroEval |
| EuroEval Polish | Multilingual | 57.1% | 73.5 | — | — | EuroEval |
| EuroEval Dutch | Multilingual | 56.2% | 72.5 | — | — | EuroEval |
| SWE-bench Pro | Coding | 55.6% | 57.8 | — | — | Xiang Deng et al. |
| IOI v1 | Coding | 54.8% | 71.1 | — | 9 Aug 2026 | Vals AI |
| EuroEval Spanish | Multilingual | 53.5% | 69.1 | — | — | EuroEval |
| Vibe Code Bench v1.1 | Coding | 53.5% | 64.4 | OpenHands | 21 Sept 2026 | Vals AI |
| ARC-AGI-2 (semi-private) | Reasoning | 52.9% | 66.4 | xhigh effort | — | ARC Prize Foundation |
| Terminal-Bench 2.0 | Agentic | 51.7% | 59.6 | — | 4 Jun 2026 | Vals AI |
| LiveBench Agentic Coding | Agentic | 50.3% | 64.0 | high effort | 25 Jun 2026 | LiveBench |
| EuroEval German | Multilingual | 49.6% | 64.2 | — | — | EuroEval |
| Chess Puzzles | Reasoning | 49.0% | 86.9 | xhigh effort | — | Epoch AI |
| OSWorld-Verified | Agentic | 47.3% | 37.7 | — | — | Tianbao Xie et al. |
| Gert Labs Composite Game Benchmark | Agentic | 46.5% | 53.8 | — | — | Gert Labs |
| Artificial Analysis Omniscience Accuracy | Knowledge | 44.3% | 68.7 | — | — | Artificial Analysis |
| FrontierMath-2025-02-28-Private | Math | 40.7% | 68.6 | xhigh effort | — | Epoch AI |
| Furniture Assembly | Reasoning | 38.3% | 66.8 | xhigh effort | — | Epoch AI |
| Artificial Analysis Humanity's Last Exam | Knowledge | 37.7% | 65.4 | — | — | Artificial Analysis |
| SimpleQA Verified | Knowledge | 37.1% | 55.9 | xhigh effort | — | Epoch AI |
| JobBench | Agentic | 34.3% | 57.7 | — | — | Yuetai Li et al. |
| τ²-bench Banking | Agentic | 32.2% | 22.0 | high effort · Sierra | 4 Aug 2026 | Sierra Research |
| FrontierMath-Tier-4-v2-Private | Math | 31.7% | 63.2 | xhigh effort | — | Epoch AI |
| Artificial Analysis Intelligence Index | Knowledge | 30.4% | 60.5 | — | — | Artificial Analysis |
| EBR-bench | Reasoning | 23.0% | 65.1 | xhigh effort | — | Epoch AI |
| Mystery Game Puzzles | Reasoning | 23.0% | 58.4 | high effort | — | Epoch AI |
| FrontierMath-Tier-4-2025-07-01-Private | Math | 18.8% | 63.9 | xhigh effort | — | Epoch AI |
| Critical Physics Tasks | Reasoning | 11.6% | 62.7 | — | — | Artificial Analysis |
44 benchmarks count, from 62 of 67 results. A grey row does not count. Too few models took that benchmark.