MBPP
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
MBPP+
69.3Accuracy
13
MBPP (test)
145.99Throughput (tokens/s)
13
MBPP AceCoder-87K
52.4Accuracy
13
MBPP-ET
73Acc
13
MBPP-CN v1 (test)
53.2Pass@1
13
MBPP
5.26Accepted Length (τ)
12
MBPP+
75.66Accuracy
12
MBPP
3,511Response Length (Tokens)
12
MBPP
57.2Accuracy
12
MBPP
60Accuracy (%)
12
MBPP base
80.7Pass@1
12
MBPP++
66.14Pass@1
12
MBPP
4.78Speedup
12
MBPP
7.68Speedup
12
MBPP
1.7Latency (s)
12
MBPP
93.2Pass@1
12
MBPP
49.23Accuracy (%)
12
MBPP
42.6Accuracy
12
MBPP+ Advanced persona
95Win-rate
12
MBPP+ Researcher persona
97Win Rate
12
MBPP+ Intermediate persona
90Win-rate
12
MBPP+ Beginner persona
94Win-rate
12
MBPP
88.4ASR
12
MBPP+ strict evaluation
0.468Accuracy
12
MBPP 3-shot max_gen_len=256
51.4Accuracy
12