List your company

Agentic Action Prediction on ALFWorld in domain

95Success RateGRPO
81.4884.9988.592.01Jun 25, 2026

Evaluation results

MethodLinks
GRPOModel=Qwen3-8B2026.0695
oursModel=Qwen3-8B2026.0695
oursModel=Qwen3-4B2026.0683
GRPOModel=Qwen3-4B2026.0682