List your company

Code Reasoning on MBPP base and extended (out-of-distribution)

55.83AccuracyInftyThink+
47.759649.854851.9554.0452Feb 6, 2026

Evaluation results

MethodLinks
InftyThink+RL_Setting=Task reward only (✓ T), Sampling_Times=32, Temperature=0.72026.0255.8347.389.2749.19
InftyThink+RL_Setting=Task and efficiency reward (✓ T+E), Sampling_Times=32, Temperature=0.72026.0253.8545.924.6423.62
VanillaRL_Setting=Task reward only (✓ T), Sampling_Times=32, Temperature=0.72026.0250.0945.436.6980.36
InftyThink+RL_Setting=Cold start only (✗), Sampling_Times=32, Temperature=0.72026.0249.2842.054.7326.36
VanillaRL_Setting=Cold start only (✗), Sampling_Times=32, Temperature=0.72026.0248.0741.015.4862.45