List your company

Code Generation on MBPP-Hard

62.56Pass Rateo4-mini
18.110429.650241.1952.7298Jan 2, 2026

Evaluation results

MethodLinks
o4-miniProcessing=Postprocessed2026.0162.5633.334.11——
Claude 3.7 SonnetProcessing=Postprocessed2026.0156.6240.642.74——
GPT-4.1-MiniProcessing=Postprocessed2026.0155.2538.815.94——
Gemini-2.0-FlashProcessing=Postprocessed2026.0149.3245.665.02——
o4-miniProcessing=Filtered2026.0147.7549.552.7——
Claude 3.7 SonnetProcessing=Filtered2026.0145.0552.252.7——
Gemini-2.0-FlashProcessing=Filtered2026.0144.5952.253.15——
GPT-4.1-MiniProcessing=Filtered2026.0144.1450.455.41——
Qwen2.5-7b-InstructProcessing=Postprocessed2026.0130.1463.016.85——
Qwen2.5-3b-CoderProcessing=Postprocessed2026.0124.253.8821.92——
Qwen2.5-3b-CoderProcessing=Filtered2026.0121.6254.9523.42——
Qwen2.5-7b-InstructProcessing=Filtered2026.0119.8272.527.66——
TextBFGSKnowledge Base=None2026.01———85.4748.72
TextBFGSRetrieval=Gradient-based, Knowledge Base=MBPP KB2026.01———94.0274.36
TextBFGS-REMORetrieval=Input-based, Knowledge Base=MBPP KB2026.01———95.7378.63
TextGrad2026.01———85.4748.72
TextGrad-Momentum2026.01———88.8958.12