List your company

Embodied Household Interaction on ALFWorld Seen

81.29Success RateGRPO
4.17424.194544.21564.2355Nov 18, 2025

Evaluation results

MethodLinks
GRPOFramework=Agent-R1, Backbone=Qwen3-4B, Type=RL-optimized2025.1181.29
RLOOFramework=Agent-R1, Backbone=Qwen3-4B, Type=RL-optimized2025.1179.08
PPOFramework=Agent-R1, Backbone=Qwen3-4B, Type=RL-optimized2025.1176.42
Reinforce++Framework=Agent-R1, Backbone=Qwen3-4B, Type=RL-optimized2025.1173.84
ReActFramework=Agent-R1, Backbone=Qwen3-4B, Type=Training-free baseline2025.117.14