Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Algorithm Synthesis on Rule-based benchmark 200 contamination-free problems
Loading...
77
Correct
AlgoSkill
68.68
70.84
73
75.16
Jun 29, 2026
Correct
T-opt
S-opt
Updated 26d ago
Evaluation Results
Method
Method
Links
Correct
T-opt
S-opt
AlgoSkill
Backbone=Haiku
2026.06
77
89
100
Direct LLM
Backbone=Haiku
2026.06
75
45
95
CoT
Backbone=Haiku
2026.06
73
68
99
Direct LLM
Backbone=Llama 8B
2026.06
69
63
87
Feedback
Search any
task
Search any
task