Anthropic
Claude Sonnet 4.5 benchmark results
41 results on 11 benchmarks from 5 sources.
Also listed as: claude-sonnet-4-5-20250929-high-32k · claude-sonnet-4-5-20250929 · claude-sonnet-4-5-20250929_16K · Claude Sonnet 4.5 (16k thinking) · claude-sonnet-4-5-20250929_59K · Claude Sonnet 4.5 (59k thinking) · claude-sonnet-4-5-20250929_32K · Claude Sonnet 4.5 (32k thinking) · Claude Sonnet 4.5 (no thinking) · claude-sonnet-4-5-20250929_unknown
All benchmark results
| Benchmarks | Score | Rank | Setting | Source type | Date |
|---|---|---|---|---|---|
| LMArena Text | 1456 | 64 | Overall, style control on; 83973 votes; 95% CI 1454–1459“claude-sonnet-4-5-20250929-high-32k” | Official leaderboard | Oct 8, 2026 |
| LMArena Text | 1455 | 69 | Overall, style control on; 82735 votes; 95% CI 1452–1458“claude-sonnet-4-5-20250929” | Official leaderboard | Oct 8, 2026 |
| LMArena WebDev | 1393 | 94 | Overall; 13256 votes; 95% CI 1386–1401“claude-sonnet-4-5-20250929-high-32k” | Official leaderboard | Oct 8, 2026 |
| LMArena WebDev | 1385 | 98 | Overall; 15723 votes; 95% CI 1377–1392“claude-sonnet-4-5-20250929” | Official leaderboard | Oct 8, 2026 |
| GPQA Diamond | 82.3% | 104 | Epoch AI run, accuracy; effort: 59k thinking | Independent run | Oct 28, 2025 |
| GPQA Diamond | 81.7% | 109 | Epoch AI run, accuracy; effort: 32k thinking | Independent run | Oct 21, 2025 |
| GPQA Diamond | 78.8% | 121 | Epoch AI run, accuracy; effort: 16k thinking | Independent run | Oct 28, 2025 |
| GPQA Diamond | 73.7% | 146 | Epoch AI run, accuracy; effort: no thinking | Independent run | Sep 29, 2025 |
| SWE-bench Verified | 71.3% | 24 | Epoch AI run (Epoch scaffold); effort: no thinking | Independent run | Feb 5, 2026 |
| SWE-bench Verified | 74.8% | 9 | agent: Sonar Foundation Agent“claude-sonnet-4-5” | Official leaderboard | Nov 3, 2025 |
| SWE-bench Verified | 71.4% | 21 | agent: mini-SWE-agent“claude-sonnet-4-5-20250929” | Official leaderboard | Feb 17, 2026 |
| AIME (OTIS Mock AIME 2024–2025) | 77.8% | 115 | Epoch AI run, OTIS Mock AIME 2024–2025; effort: 59k thinking | Independent run | Oct 28, 2025 |
| AIME (OTIS Mock AIME 2024–2025) | 77.8% | 115 | Epoch AI run, OTIS Mock AIME 2024–2025; effort: 32k thinking | Independent run | Oct 21, 2025 |
| AIME (OTIS Mock AIME 2024–2025) | 71.1% | 126 | Epoch AI run, OTIS Mock AIME 2024–2025; effort: 16k thinking | Independent run | Oct 28, 2025 |
| AIME (OTIS Mock AIME 2024–2025) | 35.6% | 187 | Epoch AI run, OTIS Mock AIME 2024–2025; effort: no thinking | Independent run | Sep 29, 2025 |
| MATH Level 5 | 97.7% | 6 | Epoch AI run, accuracy; effort: 32k thinking | Independent run | Oct 21, 2025 |
| Humanity's Last Exam | 13.7% | 29 | Compiled by Epoch AI; effort: unknown thinking | Third-party compilation via Epoch AI | Sep 29, 2025 |
| OSWorld (Verified) | 62.9% | 3 | OSWorld; agent: claude-sonnet-4-5-20250929 (100 steps); effort: no thinking | Official leaderboard via Epoch AI | Oct 30, 2025 |
| OSWorld (Verified) | 58.1% | 4 | OSWorld; agent: claude-sonnet-4-5-20250929 (50 steps); effort: no thinking | Official leaderboard via Epoch AI | Oct 30, 2025 |
| OSWorld (Verified) | 42.9% | 6 | OSWorld; agent: claude-sonnet-4-5-20250929 (15 steps); effort: no thinking | Official leaderboard via Epoch AI | Oct 30, 2025 |
| τ²-bench | 72.0% | 8 | airline domain, pass^1; effort: enabled; submitted by Sierra; v1.0.1 | Official leaderboard | Feb 24, 2026 |
| τ²-bench | 70.0% | 10 | airline domain, pass^1; submitted by Anthropic; v0.1.3“Claude-Sonnet-4.5” | Vendor-reported | Sep 29, 2025 |
| τ²-bench | 25.3% | 18 | banking_knowledge domain, pass^1; effort: enabled; submitted by Sierra; v1.0.1 | Official leaderboard | Feb 24, 2026 |
| τ²-bench | 86.2% | 1 | retail domain, pass^1; submitted by Anthropic; v0.1.3“Claude-Sonnet-4.5” | Vendor-reported | Sep 29, 2025 |
| τ²-bench | 72.4% | 21 | retail domain, pass^1; effort: enabled; submitted by Sierra; v1.0.1 | Official leaderboard | Feb 24, 2026 |
| τ²-bench | 98.0% | 2 | telecom domain, pass^1; submitted by Anthropic; v0.1.3“Claude-Sonnet-4.5” | Vendor-reported | Sep 29, 2025 |
| τ²-bench | 84.9% | 14 | telecom domain, pass^1; effort: enabled; submitted by Sierra; v1.0.1 | Official leaderboard | Feb 24, 2026 |
| Terminal-Bench 2.0 | 42.7% | 4 | Terminal-Bench 2.0; agent: MAYA; effort: unknown thinking | Third-party compilation via Epoch AI | Jan 4, 2026 |
| Terminal-Bench 2.0 | 46.5% | 76 | Terminal-Bench 2.0; agent: CAMEL-AI; effort: unknown thinking | Official leaderboard via Epoch AI | Sep 29, 2025 |
| Terminal-Bench 2.0 | 43.1% | 85 | Terminal-Bench 2.0; agent: Goose; effort: unknown thinking | Official leaderboard via Epoch AI | Sep 29, 2025 |
| Terminal-Bench 2.0 | 42.8% | 89 | Terminal-Bench 2.0; agent: Terminus 2; effort: unknown thinking | Official leaderboard via Epoch AI | Sep 29, 2025 |
| Terminal-Bench 2.0 | 42.8% | 90 | Terminal-Bench 2.0; agent: Terminus 2 ; effort: no thinking | Official leaderboard via Epoch AI | Oct 31, 2025 |
| Terminal-Bench 2.0 | 42.7% | 91 | Terminal-Bench 2.0; agent: MAYA-V2; effort: unknown thinking | Official leaderboard via Epoch AI | Sep 29, 2025 |
| Terminal-Bench 2.0 | 42.6% | 93 | Terminal-Bench 2.0; agent: OpenHands ; effort: no thinking | Official leaderboard via Epoch AI | Nov 2, 2025 |
| Terminal-Bench 2.0 | 42.6% | 94 | Terminal-Bench 2.0; agent: OpenHands; effort: unknown thinking | Official leaderboard via Epoch AI | Sep 29, 2025 |
| Terminal-Bench 2.0 | 42.5% | 95 | Terminal-Bench 2.0; agent: Mini-SWE-Agent; effort: unknown thinking | Official leaderboard via Epoch AI | Sep 29, 2025 |
| Terminal-Bench 2.0 | 42.5% | 96 | Terminal-Bench 2.0; agent: Mini-SWE-Agent ; effort: no thinking | Official leaderboard via Epoch AI | Nov 3, 2025 |
| Terminal-Bench 2.0 | 40.1% | 100 | Terminal-Bench 2.0; agent: Claude Code ; effort: no thinking | Official leaderboard via Epoch AI | Nov 4, 2025 |
| Terminal-Bench 2.0 | 40.1% | 101 | Terminal-Bench 2.0; agent: Claude Code; effort: unknown thinking | Official leaderboard via Epoch AI | Sep 29, 2025 |
| BFCL v4 (function calling) | 73.2% | 2 | Overall accuracy; native function calling (FC)“Claude-Sonnet-4-5-20250929 (FC)” | Official leaderboard | Oct 17, 2026 |
| BFCL v4 (function calling) | 24.9% | 89 | Overall accuracy; prompt-based“Claude-Sonnet-4-5-20250929 (Prompt)” | Official leaderboard | Oct 17, 2026 |
FAQ
Where do these Claude Sonnet 4.5 scores come from?
From Epoch AI Benchmarking Hub, Berkeley Function Calling Leaderboard (BFCL), LMArena leaderboard dataset, SWE-bench leaderboards (swebench.com), and τ²-bench leaderboard submissions. Each row links to the original leaderboard, dataset or report, and vendor-reported numbers are labelled.