List your company

Web Navigation on WebShop (test)

91.37Score3SPO
-3.62484821.03727645.699470.361524Feb 24, 2026Mar 13, 2026Mar 30, 2026Apr 17, 2026May 4, 2026May 21, 2026Jun 8, 2026

Evaluation results

MethodLinks
3SPOModel=Qwen2.5-1.5B-Instruct, Type=RL Training2026.0691.3779.46
3SPOModel=Qwen2.5-7B-Instruct, Type=RL Training2026.0690.2880.45
MAGETeacher=Sonnet, n=52026.0590.284
MAGETeacher=Opus, n=52026.0589.782
HGPOModel=Qwen2.5-7B-Instruct, Type=RL Training2026.0688.9678.51
GiGPOModel=Qwen2.5-7B-Instruct, Type=RL Training2026.0688.9377.6
RetroAgentmode=RL-trained reflection2026.0588.982.3
RetroAgentmode=in-context, reference=Qwen-7B2026.0587.678.9
HGPOModel=Qwen2.5-1.5B-Instruct, Type=RL Training2026.0685.5671.54
GiGPOModel=Qwen2.5-1.5B-Instruct, Type=RL Training2026.0684.9566.53
PPO (with critic)Model=Qwen2.5-7B-Instruct, Type=RL Training2026.0681.468.7
RLOOModel=Qwen2.5-7B-Instruct, Type=RL Training2026.0680.365.7
GRPOModel=Qwen2.5-7B-Instruct, Type=RL Training2026.0679.366.1
GRPOModel=Qwen2.5-1.5B-Instruct, Type=RL Training2026.0675.856.8
RLOOModel=Qwen2.5-1.5B-Instruct, Type=RL Training2026.0673.952.1
PPO (with critic)Model=Qwen2.5-1.5B-Instruct, Type=RL Training2026.0673.851.5
ReflexionModel=Qwen2.5-7B-Instruct, Type=Prompting2026.0658.128.8
ReflexionModel=Qwen2.5-1.5B-Instruct, Type=Prompting2026.0655.821.9
ReActModel=Qwen2.5-7B-Instruct, Type=Prompting2026.0646.219.5
Gemini-2.5-ProModel=Closed, Type=Prompting2026.0642.535.9
ReActModel=Qwen2.5-1.5B-Instruct, Type=Prompting2026.0640.111.3
GPT-4oModel=Closed, Type=Prompting2026.0631.823.7
Qwen2.5Model=Qwen2.5-7B-Instruct, Type=Prompting2026.0626.47.8
Qwen2.5Model=Qwen2.5-1.5B-Instruct, Type=Prompting2026.0623.15.2
SELAURBase Model=Qwen2.5-7B2026.020.893579.68
GiGPOBase Model=Qwen2.5-7B2026.020.886779.29
SELAURBase Model=Qwen2.5-1.5B2026.020.881276.56
RLOOBase Model=Qwen2.5-7B2026.020.876875.78
PPOBase Model=Qwen2.5-7B2026.020.868971.87
GRPOBase Model=Qwen2.5-1.5B2026.020.835967.18
GRPOBase Model=Qwen2.5-7B2026.020.818475.78
RLOOBase Model=Qwen2.5-1.5B2026.020.802263.94
GiGPOBase Model=Qwen2.5-1.5B2026.020.801767.57
PPOBase Model=Qwen2.5-1.5B2026.020.712649.21
promptingBase Model=Qwen2.5-1.5B2026.020.2184.29
promptingBase Model=Qwen2.5-7B2026.020.02880.39