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

MBPP

Benchmarks

Task NameDataset NameSOTA ResultTrend
Code GenerationMBPP (test)
Pass@195.1
405
Code GenerationMBPP+
Pass@184.39
238
Code GenerationMBPP+
Accuracy85.75
236
Code GenerationMBPP
Pass@192.4
211
Code GenerationMBPP
Pass@189.1
193
Code GenerationMBPP
Accuracy79.8
165
Code GenerationMBPP
Accuracy (%)92.2
146
CodingMBPP
Accuracy98.4
145
CodingMBPP+
Pass@197.88
117
Code GenerationMBPP-ET
Pass@191.8
91
Code GenerationMBPP
Accuracy96.6
90
Code GenerationMBPP
Accuracy90.5
89
Code GenerationMBPP Plus (test)
Accuracy83.6
89
Code GeneratingMBPP
Pass@183.1
88
Code GenerationMBPP
Speedup7.68
79
CodingMBPP
Pass@1 Accuracy95.33
78
CodeMBPP
Pass@191.05
73
Code GenerationMBPP Code
Performance (%)83
60
Code GenerationMBPP
Pass@1 Accuracy94.2
59
Code generationMBPP
Pass@180.4
58
Function-level Code GenerationMBPP+ augmented (test)
Pass@179.6
56
Code GenerationMBPP Sanitized
Accuracy85.7
51
Code GenerationMBPP
TPS4,290
50
Code GenerationMBPP+
Score94.2
43
Code GenerationMBPP+
Pass@173.75
40
Showing 25 of 274 rows
...