List your company

Reasoning on Reasoning Gym Puzzle24

60.67Success RateCBRL GRPO
0.3516.0131.6747.33Mar 19, 2026

Evaluation results

MethodLinks
CBRL GRPOBackbone Model=Qwen2.5-3B-Instruct2026.0360.67
GRPOBackbone Model=Qwen2.5-3B-Instruct2026.0348
CBRL GRPOBackbone Model=Llama-3.2-3B-Instruct2026.0322
GRPOBackbone Model=Llama-3.2-3B-Instruct2026.0320.67
Baseline w/ Few ShotsBackbone Model=Qwen2.5-3B-Instruct2026.036.67
BaselineBackbone Model=Qwen2.5-3B-Instruct2026.034.33
Baseline w/ Few ShotsBackbone Model=Llama-3.2-3B-Instruct2026.034
BaselineBackbone Model=Llama-3.2-3B-Instruct2026.032.67