List your company

Program Repair on Real misspecified programs neutral briefs

100Repair Success RateGemini-2.5-flash
-4235077Jun 30, 2026

Evaluation results

MethodLinks
Gemini-2.5-flashFeedback regime=judge2026.06100
Gemini-2.5-flashFeedback regime=diag2026.06100
Claude 4.6Feedback regime=judge2026.06100
Claude 4.6Feedback regime=diag2026.06100
Gemini-3.1-proFeedback regime=none2026.06100
Gemini-3.1-proFeedback regime=summ.2026.06100
GPT-4.1Feedback regime=diag2026.0693
GPT-5.1Feedback regime=diag2026.0692
DeepSeek-V4-flashFeedback regime=diag2026.0690
pooled‡Feedback regime=diag2026.0684
GPT-5.1Feedback regime=check.2026.0677
Gemini-2.5-proFeedback regime=test2026.0675
Gemini-2.5-proFeedback regime=summ.2026.0675
Gemini-2.5-proFeedback regime=check.2026.0675
Gemini-2.5-proFeedback regime=judge2026.0675
Gemini-2.5-proFeedback regime=diag2026.0675
Gemini-3.1-proFeedback regime=judge2026.0671
Gemini-3.1-proFeedback regime=diag2026.0671
pooled‡Feedback regime=judge2026.0671
Gemini-2.5-flashFeedback regime=summ.2026.0667
Gemini-2.5-flashFeedback regime=check.2026.0667
DeepSeek-V4-proFeedback regime=judge2026.0667
DeepSeek-V4-proFeedback regime=diag2026.0667
GPT-4.1Feedback regime=judge2026.0664
pooled‡Feedback regime=check.2026.0663
DeepSeek-V4-flashFeedback regime=check.2026.0660
DeepSeek-V4-flashFeedback regime=judge2026.0660
GPT-4.1Feedback regime=check.2026.0657
Gemini-3.1-proFeedback regime=check.2026.0657
GPT-5.1Feedback regime=none2026.0654
Claude 4.6Feedback regime=none2026.0650
Claude 4.6Feedback regime=summ.2026.0650
Claude 4.6Feedback regime=check.2026.0650
DeepSeek-V4-flashFeedback regime=none2026.0650
Gemini-2.5-proFeedback regime=none2026.0650
GPT-5.1Feedback regime=test2026.0646
GPT-5.1Feedback regime=summ.2026.0646
GPT-5.1Feedback regime=judge2026.0646
pooled‡Feedback regime=none2026.0646
pooled‡Feedback regime=summ.2026.0646
DeepSeek-V4-proFeedback regime=none2026.0644
GPT-4o-miniFeedback regime=none2026.0641
GPT-4.1Feedback regime=none2026.0636
pooled‡Feedback regime=test2026.0636
Gemini-2.5-flashFeedback regime=none2026.0633
Gemini-2.5-flashFeedback regime=test2026.0633
DeepSeek-V4-proFeedback regime=check.2026.0633
DeepSeek-V4-flashFeedback regime=summ.2026.0630
GPT-4o-miniFeedback regime=diag2026.0618
GPT-4.1Feedback regime=summ.2026.0614
GPT-4o-miniFeedback regime=test2026.0612
GPT-4o-miniFeedback regime=judge2026.0612
DeepSeek-V4-flashFeedback regime=test2026.0610
GPT-4o-miniFeedback regime=summ.2026.060
GPT-4o-miniFeedback regime=check.2026.060
GPT-4.1Feedback regime=test2026.060
Claude 4.6Feedback regime=test2026.060
DeepSeek-V4-proFeedback regime=test2026.060
DeepSeek-V4-proFeedback regime=summ.2026.060
Gemini-3.1-proFeedback regime=test2026.060