Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Numeral Evaluation on The benchmark held-out slice
Loading...
100
Accuracy
Python Solver Oracle
95.736
96.843
97.95
99.057
Jun 20, 2026
Accuracy
Updated 1mo ago
Evaluation Results
Method
Method
Links
Accuracy
Python Solver Oracle
type=reverse-engineere...
2026.06
100
Solver-Grounded Synthetic CoT (Model)
adapter=fine-tuned ran...
2026.06
100
Base Model
2026.06
95.9
Feedback
Search any
task
Search any
task