Home Model series Claude Opus Claude Opus 4.5 Claude Opus 4.5 · SWE-bench Multilingual · 300 problems, 9 programming languages · Scores and configurations Claude Opus 4.5 · SWE-bench Multilingual · 300 problems, 9 programming languages · Scores and configurations Claude Opus 4.5 record for SWE-bench Multilingual · 300 problems, 9 programming languages. 2 score records for this release, including 2 exact values (may include historical records). Includes 2 source locations.
Scores and evaluation conditions Claude Opus 4.5 · SWE-bench Multilingual scores and configurations Metric and score Configuration and evaluation conditions Sources Pass rate 76.2% Reported 76.20% · Unit % Published Unconfirmed
默认 High Settings 64K thinking budget、200K context、默认 high、默认 sampling、5 次独立试验。 Unknowns none recorded Tools Tool access not recorded Reasoning budget high · Thinking budget 64K Framework not recorded Protocol SWE-bench Multilingual, 300 problems in 9 programming languages; average over 5 trials. Conditions 64K thinking budget; 200K context and extended thinking turned off apply to the no-thinking evaluation description, not this row. View this score in rankings Pass rate 76.2% Reported 76.20% · Unit % Published Unconfirmed
SWE-bench 无 thinking budget Settings 发布材料明确说明 SWE-bench Verified 不使用 thinking budget。 Unknowns 除发布表和脚注外未提取完整测试脚手架细节。 Tools Tool access not recorded Reasoning budget not recorded Framework not recorded Protocol SWE-bench Multilingual, 300 problems in 9 programming languages; average over 5 trials. Conditions Extended thinking turned off; 200K context. View this score in rankings
Benchmark versions and metrics Aliases SWE-Bench Multilingual
Pass rate · % · Higher is better