← SWE-bench leaderboards

Multilingual

SWE-bench · 13 models

RankModelScore
1Gemini 3 Flash72.7
2Claude 4.6 Opus72
3Claude 4.5 Opus70.7
4GLM 569.7
5Gemini 3 Pro68.7
6MiniMax M2.568.3
7Kimi K2.567.3
8Claude 4.5 Sonnet67
9GPT 5.2 (high)66.7
10GPT 5.2 Codex66.3
11Claude 4.5 Haiku64.7
12DeepSeek V3.259
13GPT 5 mini39.7