Anthropic
availableShows if the model has enough results for an index.Claude Opus 4.6
Claude Opus 4.6 is a non-reasoning model from Anthropic. 40 benchmarks count toward its score, in 7 categories.
IndexOverall score. 50 is the middle.58.9 ±3.1
CoverageShare of the index weight with results.95%
SpeedOutput tokens per second.33/s
Input / 1MUS dollars per 1M input tokens.$5 batch $2.5
Output / 1MUS dollars per 1M output tokens.$25 batch $12.5 US dollars per 1M output tokens in a batch.
ContextMaximum tokens in one request.1M
EloLMArena rating and rank.1498 (#5)
50 is the middle of the board. The range shows the doubt in the index. Batch work costs less.
75,878 votes. Elo shows what people prefer. It does not change the score.
CapabilitiesScore per category. 50 is the middle.
50 is the middleResults
40 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. |
|---|---|---|---|---|---|---|
| AIME25 first-party comparison snapshot | Math | 99.8% | — | — | — | Arcee AI |
| SuperGPQA: Scaling LLM Evaluation Across 285 Graduate Disciplines | Knowledge | 95.0% | 76.2 | — | — | Xiaoxuan Du et al. |
| Graduate-Level Google-Proof Q&A | Knowledge | 91.3% | 62.5 | — | — | David Rein et al. |
| OTIS Mock AIME 2024-2025 | Math | 91.1% | 62.5 | max effort | — | Epoch AI |
| GPQA Diamond | Knowledge | 89.2% | 60.6 | — | — | David Rein et al. |
| MMLU-Pro first-party comparison snapshot | Knowledge | 89.1% | 60.9 | — | — | Arcee AI |
| GPQA diamond | Knowledge | 88.4% | 59.8 | max effort | — | Epoch AI |
| τ²-Bench Tool-Agent-User Evaluation | Agentic | 84.8% | 59.7 | — | — | Victor Barres et al. |
| React Native Evals | Coding | 84.1% | 65.1 | — | — | Callstack |
| Artificial Analysis GPQA Diamond | Knowledge | 84.0% | 54.6 | — | — | Artificial Analysis |
| BrowseComp | Agentic | 83.7% | 70.1 | — | — | OpenAI |
| ScreenSpot Pro | Multimodal | 83.1% | 65.1 | — | — | Kaixin Li et al. |
| Massive Multitask Language Understanding Professional | Knowledge | 82.0% | 49.7 | — | — | Yubo Wang et al. |
| Software Engineering Benchmark Verified | Coding | 80.8% | 62.4 | — | — | Carlos E. Jimenez et al. |
| Massive Multi-discipline Multimodal Understanding Pro | Multimodal | 77.3% | 53.1 | — | — | MMMU-Pro authors |
| SWE-Bench verified | Coding | 77.2% | 59.4 | — | — | Epoch AI |
| SWE-bench Verified (mini-swe-agent-v2) | Coding | 75.6% | 58.2 | — | — | Arcee AI |
| SWE-bench Verified | Coding | 75.6% | 58.2 | mini-SWE-agent | 1 Sept 2026 | SWE-bench team |
| DeepSearchQA | Agentic | 73.7% | 56.5 | — | — | Meta AI |
| OSWorld-Verified | Agentic | 72.7% | 61.4 | — | — | Tianbao Xie et al. |
| Artificial Analysis MMMU-Pro | Multimodal | 72.5% | 55.7 | — | — | Artificial Analysis |
| SWE-bench Multilingual | Coding | 72.0% | — | mini-SWE-agent | 2 Sept 2026 | SWE-bench team |
| LiveCodeBench Pro | Coding | 70.7% | — | — | — | LiveCodeBench Pro authors |
| Claw-Eval | Agentic | 70.4% | 69.0 | — | — | Bowen Ye et al. |
| Artificial Analysis Long Context Reasoning | Reasoning | 67.0% | 54.6 | — | — | Artificial Analysis |
| CyberGym | Agentic | 66.6% | 61.7 | — | — | Zhun Wang et al. |
| FrontierMath-Tiers-1-3-v2-Private | Math | 66.0% | 68.6 | max effort | — | Epoch AI |
| SWE-Rebench | Coding | 65.3% | — | — | — | Nebius |
| MedXpertQA Multimodal | Multimodal | 64.8% | — | — | — | Meta AI |
| Gert Labs Composite Game Benchmark | Agentic | 61.9% | 67.3 | — | — | Gert Labs |
| Vibe Code Bench v1.1 | Coding | 57.6% | 66.1 | OpenHands | 21 Sept 2026 | Vals AI |
| SWE-bench Pro | Coding | 53.4% | 55.6 | — | — | Xiang Deng et al. |
| Humanity's Last Exam | Knowledge | 53.0% | 73.7 | — | — | Center for AI Safety et al. |
| MedXpertQA Text | Knowledge | 52.1% | — | — | — | Meta AI |
| ERQA | Multimodal | 51.6% | 47.1 | — | — | Qwen |
| SimpleQA Verified | Knowledge | 47.0% | 65.1 | max effort | — | Epoch AI |
| Artificial Analysis Omniscience Accuracy | Knowledge | 45.8% | 70.5 | — | — | Artificial Analysis |
| Artificial Analysis IFBench | Instruction | 44.6% | 34.4 | — | — | Artificial Analysis |
| FrontierMath-2025-02-28-Private | Math | 40.7% | 68.6 | max effort | — | Epoch AI |
| Humanity's Last Exam without tools | Knowledge | 40.0% | 62.7 | — | — | OpenAI |
| JobBench | Agentic | 36.7% | 59.3 | — | — | Yuetai Li et al. |
| Furniture Assembly | Reasoning | 28.3% | 59.7 | max effort | — | Epoch AI |
| τ²-bench Banking | Agentic | 27.3% | 18.4 | max effort · Sierra | 4 Aug 2026 | Sierra Research |
| FrontierCode 1.1 Main | Coding | 26.9% | 57.4 | — | — | Cognition |
| FrontierMath-Tier-4-v2-Private | Math | 26.8% | 60.8 | max effort | — | Epoch AI |
| Artificial Analysis Intelligence Index | Knowledge | 26.4% | 55.4 | — | — | Artificial Analysis |
| Mystery Game Puzzles | Reasoning | 25.0% | 60.5 | max effort | — | Epoch AI |
| FrontierMath-Tier-4-2025-07-01-Private | Math | 22.9% | 68.3 | max effort | — | Epoch AI |
| ResearchClawBench | Agentic | 19.9% | — | — | — | InternScience |
| Artificial Analysis Humanity's Last Exam | Knowledge | 19.1% | 45.3 | — | — | Artificial Analysis |
| HealthBench Hard | Knowledge | 14.8% | 49.9 | — | — | Meta AI |
| Chess Puzzles | Reasoning | 14.0% | 41.6 | max effort | — | Epoch AI |
| EBR-bench | Reasoning | 12.7% | 58.0 | max effort | — | Epoch AI |
| ApprenticeBench: end-to-end computer use, continual learning, and long-horizon agency on a real accounts-payable job | Agentic | 5.0% | 62.9 | — | — | NeoCognition |
| Critical Physics Tasks | Reasoning | 2.8% | 44.2 | — | — | Artificial Analysis |
40 benchmarks count, from 48 of 55 results. A grey row does not count. Too few models took that benchmark.