List your company

Premise Selection on Isabelle (IS) (test)

86.3Exact Match AccuracyLlama-3.1 (RULES)
50.21259.58168.9578.319Feb 1, 2026

Evaluation results

MethodLinks
Llama-3.1 (RULES)Optimization=RULES2026.0286.3
Llama-3.1 (GRPO)Optimization=GRPO2026.0285.9
Qwen2.5 (RULES)Optimization=RULES2026.0281.3
OLMO2 (RULES)Optimization=RULES2026.0276.6
DeepMath (RULES)Optimization=RULES2026.0275.8
Qwen2.5 (GRPO)Optimization=GRPO2026.0274.2
OLMO2 (GRPO)Optimization=GRPO2026.0272.2
DeepMath (GRPO)Optimization=GRPO2026.0267.6
Qwen2.5 (vanilla)Optimization=vanilla2026.0255.5
Llama-3.1 (vanilla)Optimization=vanilla2026.0254.7
OLMO2 (vanilla)Optimization=vanilla2026.0253.9
DeepMath (vanilla)Optimization=vanilla2026.0251.6