Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Code Generation on MBPP+ (Accuracy, Time, Token)
Loading...
82.8
Accuracy
LaTER (training)
27.16
41.605
56.05
70.495
Apr 28, 2026
Apr 29, 2026
May 1, 2026
May 3, 2026
May 4, 2026
May 6, 2026
May 8, 2026
Accuracy
Execution Time
Token Count
Updated 23d ago
Evaluation Results
Method
Method
Links
Accuracy
Execution Time
Token Count
LaTER (training)
Backbone=Qwen3-14B
2026.05
82.8
-
1,717
CoT Baseline
Backbone=Qwen3-14B
2026.05
82
-
2,144
LaTER (training-free)
Backbone=Qwen3-14B
2026.05
79.6
-
1,760
CoT-SFT
Backbone=Qwen3-14B
2026.05
78
-
1,817
RecursiveMAS
Recursion Round=r=3, S...
2026.04
37.4
805
595
RecursiveMAS
Recursion Round=r=2, S...
2026.04
36.9
627
531
RecursiveMAS
Recursion Round=r=1, S...
2026.04
35.1
449
577
Recursive-TextMAS
Recursion Round=r=1, S...
2026.04
30.7
976
1,146
Recursive-TextMAS
Recursion Round=r=2, S...
2026.04
30
1,847
1,998
Recursive-TextMAS
Recursion Round=r=3, S...
2026.04
29.3
2,310
2,676
Feedback
Search any
task
Search any
task