context.vn

swe Atlas Refactoring

10 models evaluated

#ModelProviderTypeScore
2GPT-5.5OpenAIClosedGpt-5.5 (Codex)
3GPT-5.4OpenAIClosedGpt-5.4 (Codex)
4GPT-5.3 CodexGpt-5.3 (Codex)OpenAIClosed
5Claude Opus 4.6Opus-4.6 (Claude Code)AnthropicClosed
6Gemini 3.1 ProGemini-3.1-Pro (Gemini CLI)GoogleClosed
7Claude Sonnet 4.6Sonnet-4.6 (Claude Code)AnthropicClosed
8GLM-5Glm-5 (Mini-SWE-Agent)Z.AIOpen
9Kimi K2.5Kimi-K2.5 (Mini-SWE-Agent)Moonshot AIOpen
10MiniMax M2.5Minimax-M2.5 (Mini-SWE-Agent)MiniMaxClosed
11Gemini 3 FlashGemini-3-Flash (Mini-SWE-Agent)GoogleClosed