List your company

Web Navigation on WebShop unseen (test)

87.2ScoreProxMO
22.61639.38356.1572.917Feb 22, 2026

Evaluation results

MethodLinks
ProxMOBackbone=Qwen2.5-7B-Instruct2026.0287.276.5
GiGPOBackbone=Qwen2.5-7B-Instruct2026.0285.574.8
ProxMOBackbone=Qwen2.5-1.5B-Instruct2026.0285.367.1
GiGPOBackbone=Qwen2.5-1.5B-Instruct2026.0281.762.3
GRPOBackbone=Qwen2.5-7B-Instruct2026.0279.267.2
GRPOBackbone=Qwen2.5-1.5B-Instruct2026.0273.152.2
ReflexionBackbone=Qwen2.5-1.5B-Instruct2026.0258.623.5
ReflexionBackbone=Qwen2.5-7B-Instruct2026.0256.330.2
ReActBackbone=Qwen2.5-7B-Instruct2026.0247.821
Gemini-2.5-ProBackbone=Closed-Source LLMs2026.0242.535.9
ReActBackbone=Qwen2.5-1.5B-Instruct2026.0242.114.3
GPT-4oBackbone=Closed-Source LLMs2026.0231.823.7
BaseBackbone=Qwen2.5-1.5B-Instruct2026.0225.16.3
BaseBackbone=Qwen2.5-7B-Instruct2026.0225.18.4