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

MBPP+

Benchmarks

Task NameDataset NameSOTA ResultTrend
Code GenerationMBPP+ Out-of-Domain (test)
Accuracy65.61
18
Unit test generationMBPP+ (test)
Error Rate0.53
7
Code GenerationMBPP+
Success Rate (avg@4)70.8
6
Code GenerationMBPP+ full latest
TPF670
3
Code GenerationMBPP+ 0-shot
Accuracy79.36
3
Showing 5 of 5 rows