Share your thoughts, 1 month free Claude Pro on usSee more

Program Repair on Real misspecified programs neutral briefs

100Repair Success Rate

Gemini-2.5-flash

Updated 24d ago

Evaluation Results

Method	Links
Gemini-2.5-flash 2026.06		100
Gemini-2.5-flash 2026.06		100
Claude 4.6 2026.06		100
Claude 4.6 2026.06		100
Gemini-3.1-pro 2026.06		100
Gemini-3.1-pro 2026.06		100
GPT-4.1 2026.06		93
GPT-5.1 2026.06		92
DeepSeek-V4-flash 2026.06		90
pooled‡ 2026.06		84
GPT-5.1 2026.06		77
Gemini-2.5-pro 2026.06		75
Gemini-2.5-pro 2026.06		75
Gemini-2.5-pro 2026.06		75
Gemini-2.5-pro 2026.06		75
Gemini-2.5-pro 2026.06		75
Gemini-3.1-pro 2026.06		71
Gemini-3.1-pro 2026.06		71
pooled‡ 2026.06		71
Gemini-2.5-flash 2026.06		67
Gemini-2.5-flash 2026.06		67
DeepSeek-V4-pro 2026.06		67
DeepSeek-V4-pro 2026.06		67
GPT-4.1 2026.06		64
pooled‡ 2026.06		63
DeepSeek-V4-flash 2026.06		60
DeepSeek-V4-flash 2026.06		60
GPT-4.1 2026.06		57
Gemini-3.1-pro 2026.06		57
GPT-5.1 2026.06		54
Claude 4.6 2026.06		50
Claude 4.6 2026.06		50
Claude 4.6 2026.06		50
DeepSeek-V4-flash 2026.06		50
Gemini-2.5-pro 2026.06		50
GPT-5.1 2026.06		46
GPT-5.1 2026.06		46
GPT-5.1 2026.06		46
pooled‡ 2026.06		46
pooled‡ 2026.06		46
DeepSeek-V4-pro 2026.06		44
GPT-4o-mini 2026.06		41
GPT-4.1 2026.06		36
pooled‡ 2026.06		36
Gemini-2.5-flash 2026.06		33
Gemini-2.5-flash 2026.06		33
DeepSeek-V4-pro 2026.06		33
DeepSeek-V4-flash 2026.06		30
GPT-4o-mini 2026.06		18
GPT-4.1 2026.06		14
GPT-4o-mini 2026.06		12
GPT-4o-mini 2026.06		12
DeepSeek-V4-flash 2026.06		10
GPT-4o-mini 2026.06		0
GPT-4o-mini 2026.06		0
GPT-4.1 2026.06		0
Claude 4.6 2026.06		0
DeepSeek-V4-pro 2026.06		0
DeepSeek-V4-pro 2026.06		0
Gemini-3.1-pro 2026.06		0