Anthropic
availableShows if the model has enough results for an index.Claude Sonnet 4.6
Claude Sonnet 4.6 is a non-reasoning model from Anthropic. 42 benchmarks count toward its score, in 8 categories.
IndexOverall score out of 100.58.1 ±2.5
CoverageShare of the index weight with results.100%
SpeedOutput tokens per second.38/s
Input / 1MUS dollars per 1M input tokens.$3 batch $1.5
Output / 1MUS dollars per 1M output tokens.$15 batch $7.5 US dollars per 1M output tokens in a batch.
ContextMaximum tokens in one request.1M
EloLMArena rating and rank.1458 (#35)
The index is a score out of 100. The ± range shows how much it can change. Batch work costs less.
66,208 votes. Elo shows what people prefer. It does not change the score.
CapabilitiesScore per category, out of 100.
Out of 100Results
42 counted| BenchmarkThe test name. | CategoryThe capability that the test measures. | ResultThe score from the publisher. | IndexThis result as a score out of 100. | RunThe settings of the run. | DateDate of the result. | Published byThe source of the result. |
|---|---|---|---|---|---|---|
| SuperGPQA: Scaling LLM Evaluation Across 285 Graduate Disciplines | Knowledge | 95.0% | 76.2 | — | — | Xiaoxuan Du et al. |
| AIME | Math | 92.3% | 56.7 | — | 16 Apr 2026 | Vals AI |
| Graduate-Level Google-Proof Q&A | Knowledge | 89.9% | 61.2 | — | — | David Rein et al. |
| MMLU Pro | Knowledge | 87.3% | 58.2 | — | 1 Sept 2026 | Vals AI |
| ARC-AGI-1 (semi-private) | Reasoning | 86.0% | 67.8 | max effort | — | ARC Prize Foundation |
| GPQA Diamond | Knowledge | 85.6% | 57.3 | — | 1 Sept 2026 | Vals AI |
| MMMU Pro | Multimodal | 83.6% | 63.4 | — | 1 Sept 2026 | Vals AI |
| LiveCodeBench | Coding | 82.1% | 58.9 | — | 1 Sept 2026 | Vals AI |
| React Native Evals | Coding | 80.6% | 60.3 | — | — | Callstack |
| Artificial Analysis GPQA Diamond | Knowledge | 79.9% | 50.4 | — | — | Artificial Analysis |
| Software Engineering Benchmark Verified | Coding | 79.6% | 61.4 | — | — | Carlos E. Jimenez et al. |
| τ²-Bench Tool-Agent-User Evaluation | Agentic | 79.5% | 55.9 | — | — | Victor Barres et al. |
| Massive Multitask Language Understanding Professional | Knowledge | 79.2% | 45.3 | — | — | Yubo Wang et al. |
| GPQA diamond | Knowledge | 78.8% | 51.0 | max effort | — | Epoch AI |
| CharXiv Reasoning | Multimodal | 77.4% | 54.2 | — | — | CharXiv authors |
| SWE-bench | Coding | 77.4% | 59.6 | — | 1 Sept 2026 | Vals AI |
| SWE-Bench verified | Coding | 75.2% | 57.9 | — | — | Epoch AI |
| OSWorld-Verified | Agentic | 72.1% | 60.8 | — | — | Tianbao Xie et al. |
| OTIS Mock AIME 2024-2025 | Math | 71.1% | 51.3 | max effort | — | Epoch AI |
| Artificial Analysis MMMU-Pro | Multimodal | 70.6% | 53.4 | — | — | Artificial Analysis |
| EuroEval French | Multilingual | 68.5% | 87.8 | — | — | EuroEval |
| Artificial Analysis Long Context Reasoning | Reasoning | 68.3% | 55.5 | — | — | Artificial Analysis |
| Claw-Eval | Agentic | 67.8% | 65.0 | — | — | Bowen Ye et al. |
| EuroEval Portuguese | Multilingual | 67.0% | 85.9 | — | — | EuroEval |
| EuroEval Italian | Multilingual | 65.3% | 83.8 | — | — | EuroEval |
| CyberGym | Agentic | 65.2% | 60.8 | — | — | Zhun Wang et al. |
| EuroEval Swedish | Multilingual | 64.5% | 82.8 | — | — | EuroEval |
| Gert Labs Composite Game Benchmark | Agentic | 62.9% | 68.3 | — | — | Gert Labs |
| SWE-Rebench | Coding | 60.7% | — | — | — | Nebius |
| EuroEval Polish | Multilingual | 59.9% | 77.0 | — | — | EuroEval |
| EuroEval Dutch | Multilingual | 59.7% | 76.8 | — | — | EuroEval |
| Terminal-Bench 2.0 | Agentic | 59.6% | 65.2 | — | 4 Jun 2026 | Vals AI |
| EuroEval Spanish | Multilingual | 58.5% | 75.4 | — | — | EuroEval |
| ARC-AGI-2 (semi-private) | Reasoning | 58.3% | 69.1 | max effort | — | ARC Prize Foundation |
| Terminal-Bench 2.1 | Agentic | 57.3% | 57.9 | — | 21 Sept 2026 | Vals AI |
| EuroEval German | Multilingual | 54.8% | 70.7 | — | — | EuroEval |
| Vibe Code Bench v1.1 | Coding | 51.5% | 63.6 | OpenHands | 21 Sept 2026 | Vals AI |
| SkillsBench | Coding | 49.1% | 64.6 | OpenHands | 11 Sept 2026 | Vals AI |
| Humanity's Last Exam | Knowledge | 49.0% | 70.3 | — | — | Center for AI Safety et al. |
| cursorBench31 | Coding | 48.8% | — | — | — | Benchmark authors |
| Artificial Analysis IFBench | Instruction | 41.2% | 30.9 | — | — | Artificial Analysis |
| Code Migration | Coding | 39.9% | 70.5 | — | 21 Sept 2026 | Vals AI |
| Artificial Analysis Omniscience Accuracy | Knowledge | 38.6% | 61.6 | — | — | Artificial Analysis |
| JobBench | Agentic | 36.9% | 59.5 | — | — | Yuetai Li et al. |
| SimpleQA Verified | Knowledge | 32.8% | 51.9 | max effort | — | Epoch AI |
| FrontierMath-2025-02-28-Private | Math | 32.4% | 60.9 | — | — | Epoch AI |
| Artificial Analysis Intelligence Index | Knowledge | 24.7% | 53.3 | — | — | Artificial Analysis |
| FrontierCode 1.1 Main | Coding | 24.3% | 55.0 | — | — | Cognition |
| Mystery Game Puzzles | Reasoning | 16.0% | 50.9 | low effort | — | Epoch AI |
| Artificial Analysis Humanity's Last Exam | Knowledge | 13.3% | 39.0 | — | — | Artificial Analysis |
| OSWorld 2.0 | Agentic | 8.3% | 59.9 | — | — | Mengqi Yuan et al. |
| FrontierMath-Tier-4-2025-07-01-Private | Math | 8.3% | 52.5 | — | — | Epoch AI |
| Agent Arena command recovery | Agentic | 5.0 | 73.0 | — | 15 Sept 2026 | LMArena |
| Chess Puzzles | Reasoning | 3.0% | 27.4 | max effort | — | Epoch AI |
| ApprenticeBench: end-to-end computer use, continual learning, and long-horizon agency on a real accounts-payable job | Agentic | 2.0% | 61.2 | — | — | NeoCognition |
| Critical Physics Tasks | Reasoning | 0.9% | 40.3 | — | — | Artificial Analysis |
| ProgramBench | Coding | 0.5% | — | — | 21 Sept 2026 | Vals AI |
| Agent Arena steerability | Agentic | -5.0 | 61.7 | — | 15 Sept 2026 | LMArena |
| Agent Arena task outcome | Agentic | -5.8 | 60.8 | — | 15 Sept 2026 | LMArena |
42 benchmarks count, from 56 of 59 results. A grey row does not count. Too few models took that benchmark.