Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Competitive Programming on Hard Bench (HB-15) pass@5 v2
Loading...
47
Correlation Score
Reflexion
18.92
26.21
33.5
40.79
Jun 29, 2026
Correlation Score
Time Optimization Rate
Updated 25d ago
Evaluation Results
Method
Method
Links
Correlation Score
Time Optimization Rate
Reflexion
Backbone=Haiku-4-5, Ev...
2026.06
47
86
AlgoSkill (MCTS)
Backbone=Haiku-4-5, Va...
2026.06
47
100
CoT
Backbone=Sonnet-4-5, E...
2026.06
40
83
AlgoSkill-G
Backbone=Sonnet-4-5, V...
2026.06
40
83
CoT
Backbone=Haiku, Evalua...
2026.06
27
100
AlgoSkill-G
Backbone=Haiku, Varian...
2026.06
27
100
Direct
Backbone=Sonnet-4-5, E...
2026.06
27
0
Self-Refine
Backbone=Haiku-4-5, Ev...
2026.06
27
-
Direct
Backbone=Haiku, Evalua...
2026.06
20
100
Feedback
Search any
task
Search any
task