DeepSWE leaderboard

Top 10 by Pass@1 · 21 models listed by the source

chatjudge.ai
DeepSWE: top 10 by Pass@1Bars from zero. Leader: GPT-6 Astra (xhigh) at 74.12%. Full values at chatjudge.ai.0%20%40%60%80%GPT-6 Astra (xhigh) (OpenAI), rank 1: 74.12%OGPT-6 Astra74.12%Gemini 3.8 Flash (high) (Google), rank 2: 73.83%GGemini 3.8 Flash73.83%Claude Opus 5 (max) (Anthropic), rank 3: 73.65%AClaude Opus 573.65%GPT-5.6 Sol (max) (OpenAI), rank 4: 72.67%OGPT-5.6 Sol72.67%Claude Fable 5 (xhigh) (Anthropic), rank 5: 69.91%AClaude Fable 569.91%GLM 5.3 (max) (Z.ai), rank 6: 68.96%ZGLM 5.368.96%Kimi K3 (max) (Moonshot), rank 7: 68.51%KKimi K368.51%Grok 4.6 (medium) (xAI), rank 8: 67.48%XGrok 4.667.48%GPT-5.6 Luna (max) (OpenAI), rank 9: 67.19%OGPT-5.6 Luna67.19%GPT-5.5 (xhigh) (OpenAI), rank 10: 67.04%OGPT-5.567.04%
DeepSWE: top 10 by Pass@1Bars from zero. Leader: GPT-6 Astra (xhigh) at 74.12%. Full values at chatjudge.ai.0%25%50%75%GPT-6 Astra (xhigh) (OpenAI), rank 1: 74.12%OGPT-6 Astra74.12%Gemini 3.8 Flash (high) (Google), rank 2: 73.83%GGemini 3.8 Flash73.83%Claude Opus 5 (max) (Anthropic), rank 3: 73.65%AClaude Opus 573.65%GPT-5.6 Sol (max) (OpenAI), rank 4: 72.67%OGPT-5.6 Sol72.67%Claude Fable 5 (xhigh) (Anthropic), rank 5: 69.91%AClaude Fable 569.91%GLM 5.3 (max) (Z.ai), rank 6: 68.96%ZGLM 5.368.96%Kimi K3 (max) (Moonshot), rank 7: 68.51%KKimi K368.51%Grok 4.6 (medium) (xAI), rank 8: 67.48%XGrok 4.667.48%GPT-5.6 Luna (max) (OpenAI), rank 9: 67.19%OGPT-5.6 Luna67.19%GPT-5.5 (xhigh) (OpenAI), rank 10: 67.04%OGPT-5.567.04%