Anthropic
availableShows if the model has enough results for an index.Claude Sonnet 5
Claude Sonnet 5 is a reasoning model from Anthropic. 49 benchmarks count toward its score, in 7 categories.
IndexOverall score out of 100.67.6 ±2.8
CoverageShare of the index weight with results.95%
SpeedOutput tokens per second.44/s
Input / 1MUS dollars per 1M input tokens.$2 batch $1
Output / 1MUS dollars per 1M output tokens.$10 batch $5 US dollars per 1M output tokens in a batch.
ContextMaximum tokens in one request.1M
EloLMArena rating and rank.1442 (#62)
The index is a score out of 100. The ± range shows how much it can change. Batch work costs less.
35,301 votes. Elo shows what people prefer. It does not change the score.
CapabilitiesScore per category, out of 100.
Out of 100Results
49 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. |
|---|---|---|---|---|---|---|
| LiveBench Mathematics | Math | 92.9% | 72.1 | xhigh effort | 25 Jun 2026 | LiveBench |
| Artificial Analysis GPQA Diamond | Knowledge | 91.1% | 61.9 | — | — | Artificial Analysis |
| VulcanBench Coding Intelligence Index v1 | Coding | 89.2% | — | — | — | VulcanBench contributors |
| GPQA Diamond | Knowledge | 88.9% | 60.3 | — | 1 Sept 2026 | Vals AI |
| LiveBench Reasoning | Reasoning | 88.7% | 79.2 | xhigh effort | 25 Jun 2026 | LiveBench |
| CharXiv Reasoning | Multimodal | 88.3% | 66.6 | — | — | CharXiv authors |
| MMLU Pro | Knowledge | 87.5% | 58.5 | — | 1 Sept 2026 | Vals AI |
| Software Engineering Benchmark Verified | Coding | 85.2% | 65.9 | — | — | Carlos E. Jimenez et al. |
| BrowseComp | Agentic | 84.7% | 70.9 | — | — | OpenAI |
| MMMU Pro | Multimodal | 83.0% | 62.4 | — | 1 Sept 2026 | Vals AI |
| LiveCodeBench | Coding | 82.4% | 59.3 | — | 1 Sept 2026 | Vals AI |
| Artificial Analysis Long Context Reasoning | Reasoning | 82.0% | 65.0 | — | — | Artificial Analysis |
| Vibe Code Bench v1.1 | Coding | 81.3% | 76.0 | OpenHands | 21 Sept 2026 | Vals AI |
| OSWorld-Verified | Agentic | 81.2% | 69.3 | — | — | Tianbao Xie et al. |
| LiveBench Coding | Coding | 80.7% | 71.6 | xhigh effort | 25 Jun 2026 | LiveBench |
| GPQA diamond | Knowledge | 80.3% | 52.4 | max effort | — | Epoch AI |
| LABBench2: An Improved Benchmark for AI Systems Performing Biology Research | Knowledge | 80.1% | — | — | — | Jon M. Laurent et al. |
| OTIS Mock AIME 2024-2025 | Math | 80.0% | 56.3 | max effort | — | Epoch AI |
| SWE-bench | Coding | 79.6% | 61.4 | — | 1 Sept 2026 | Vals AI |
| Artificial Analysis MMMU-Pro | Multimodal | 77.3% | 61.6 | — | — | Artificial Analysis |
| CharXiv Reasoning without tools | Multimodal | 77.0% | — | — | — | CharXiv authors |
| ProofBench v1.1 | Math | 77.0% | 80.6 | — | 21 Sept 2026 | Vals AI |
| LiveBench Language | Knowledge | 75.0% | 61.0 | xhigh effort | 25 Jun 2026 | LiveBench |
| Terminal-Bench 2.1 | Agentic | 74.5% | 68.0 | — | 21 Sept 2026 | Vals AI |
| LiveBench Data Analysis | Reasoning | 71.7% | 55.6 | xhigh effort | 25 Jun 2026 | LiveBench |
| Artificial Analysis Coding Index | Coding | 71.5% | 69.4 | — | — | Artificial Analysis |
| FrontierMath-Tiers-1-3-v2-Private | Math | 65.6% | 68.4 | max effort | — | Epoch AI |
| LiveBench Instruction Following | Instruction | 63.9% | 60.9 | xhigh effort | 25 Jun 2026 | LiveBench |
| SWE-bench Pro | Coding | 63.2% | 65.1 | — | — | Xiang Deng et al. |
| cursorBench32 | Coding | 61.5% | 69.9 | — | — | Benchmark authors |
| LiveBench Agentic Coding | Agentic | 59.4% | 72.6 | xhigh effort | 25 Jun 2026 | LiveBench |
| Humanity's Last Exam with tools | Agentic | 57.4% | 70.2 | — | — | DeepSeek-AI |
| Humanity's Last Exam | Knowledge | 57.4% | 77.4 | — | — | Center for AI Safety et al. |
| Medical Long Context Reasoning (MLCR-AA) | Reasoning | 55.0% | 85.9 | — | — | Wisedocs and Artificial Analysis |
| Artificial Analysis SciCode | Coding | 54.3% | 68.2 | — | — | Artificial Analysis |
| GDPval-AA normalized | Agentic | 47.5% | 72.0 | — | — | Artificial Analysis |
| SkillsBench | Coding | 46.5% | 62.4 | OpenHands | 11 Sept 2026 | Vals AI |
| Artificial Analysis AnalystAgent | Agentic | 46.3% | 74.0 | — | — | Artificial Analysis |
| IOI | Coding | 45.0% | 65.2 | — | 21 Sept 2026 | Vals AI |
| Code Migration | Coding | 44.4% | 73.3 | — | 21 Sept 2026 | Vals AI |
| Artificial Analysis Agentic Index | Agentic | 44.3% | 73.3 | — | — | Artificial Analysis |
| Humanity's Last Exam without tools | Knowledge | 43.2% | 65.4 | — | — | OpenAI |
| FrontierCode 1.1 Main | Coding | 42.7% | 71.5 | — | — | Cognition |
| Artificial Analysis Humanity's Last Exam | Knowledge | 41.3% | 69.3 | — | — | Artificial Analysis |
| Artificial Analysis Omniscience Accuracy | Knowledge | 40.1% | 63.5 | — | — | Artificial Analysis |
| Artificial Analysis Intelligence Index | Knowledge | 38.2% | 70.1 | — | — | Artificial Analysis |
| Mystery Game Puzzles | Reasoning | 35.0% | 71.1 | max effort | — | Epoch AI |
| cursorBench40 | Coding | 34.1% | — | — | — | Benchmark authors |
| SimpleQA Verified | Knowledge | 33.7% | 52.7 | max effort | — | Epoch AI |
| HLE-Verified | Knowledge | 31.0% | — | — | — | Weiqi Zhai et al. |
| FrontierMath-Tier-4-v2-Private | Math | 29.3% | 62.0 | max effort | — | Epoch AI |
| Critical Physics Tasks | Reasoning | 16.9% | 73.8 | — | — | Artificial Analysis |
| ApprenticeBench: end-to-end computer use, continual learning, and long-horizon agency on a real accounts-payable job | Agentic | 16.0% | 69.0 | — | — | NeoCognition |
| Chess Puzzles | Reasoning | 16.0% | 44.2 | max effort | — | Epoch AI |
| Terminal-Bench 3.0 | Agentic | 14.6% | 66.0 | — | — | Ryan Marten et al. |
| Vibe Code Bench 1-100 | Coding | 13.8% | 66.9 | OpenHands | 16 Sept 2026 | Vals AI |
| Agent Arena steerability | Agentic | 8.4 | 76.8 | high effort | 15 Sept 2026 | LMArena |
| Terminal-Bench 4.0 | Agentic | 8.1% | 65.2 | — | 21 Sept 2026 | Vals AI |
| Agent Arena command recovery | Agentic | 7.4 | 75.7 | high effort | 15 Sept 2026 | LMArena |
| Agent Arena task outcome | Agentic | 2.9 | 70.6 | high effort | 15 Sept 2026 | LMArena |
| ProgramBench | Coding | 0.0% | — | — | 21 Sept 2026 | Vals AI |
49 benchmarks count, from 55 of 61 results. A grey row does not count. Too few models took that benchmark.