Claude 3.5 Sonnet

anthropic · Ranked across 6 benchmarks · best rank #15

Benchmark scores

BenchmarkCategoryRankScoreCaptured
METR Time Horizon agents #15 21m 2024-10-22
Aider Polyglot code #18 51.6% 2025-01-17
SWE-bench Verified agents #20 62.8% 2025-02-28
AA Coding Index code #30 30.2% 2026-06-15
Chatbot Arena chat #48 1298 2026-06-10
OpenRouter · Weekly Usage usage #59 #602 2026-06-09