List your company

Agentic web navigation and interaction on WebShop (test)

84.38SRRePro
-2.959219.715442.3965.0646Jun 12, 2026

Evaluation results

MethodLinks
ReProBackbone=Qwen2.5-7B, Protocol=Best, Category=RePro (Ours)2026.0684.3891.185.475.25
ReProBackbone=Qwen2.5-3B, Protocol=Best, Category=RePro (Ours)2026.0683.5992.0911.717.68
ReProBackbone=Qwen2.5-1.5B, Protocol=Best, Category=RePro (Ours)2026.0681.6491.728.21.9
L1Backbone=Qwen2.5-7B, Protocol=Best, Category=Training Methods2026.0680.8691.741.955.81
ReProBackbone=Qwen2.5-7B, Protocol=Avg, Category=RePro (Ours)2026.0680.4789.135.824.04
ReProBackbone=Qwen2.5-3B, Protocol=Avg, Category=RePro (Ours)2026.0679.7390.7511.576.37
Meta PromptBackbone=Qwen2.5-7B, Protocol=Best, Category=Training Methods2026.0678.9185.93——
L1Backbone=Qwen2.5-3B, Protocol=Best, Category=Training Methods2026.0676.5688.534.684.12
L2Backbone=Qwen2.5-3B, Protocol=Best, Category=Training Methods2026.0676.5688.64.684.19
ReProBackbone=Qwen2.5-1.5B, Protocol=Avg, Category=RePro (Ours)2026.0676.1387.568.982.29
L1Backbone=Qwen2.5-7B, Protocol=Avg, Category=Training Methods2026.0675.6287.810.972.72
L2Backbone=Qwen2.5-7B, Protocol=Best, Category=Training Methods2026.0675.3987.54-3.521.61
L2Backbone=Qwen2.5-1.5B, Protocol=Best, Category=Training Methods2026.067588.011.56-1.81
Meta PromptBackbone=Qwen2.5-7B, Protocol=Avg, Category=Training Methods2026.0674.6585.09——
GiGPOBackbone=Qwen2.5-7B, Protocol=Best, Category=Training Methods2026.0674.2283.98-4.69-1.95
Meta PromptBackbone=Qwen2.5-1.5B, Protocol=Best, Category=Training Methods2026.0673.4489.82——
GRPOBackbone=Qwen2.5-7B, Protocol=Best, Category=Training Methods2026.0673.2982.18-5.62-3.75
GiGPOBackbone=Qwen2.5-3B, Protocol=Best, Category=Training Methods2026.0671.8883.940-0.47
Meta PromptBackbone=Qwen2.5-3B, Protocol=Best, Category=Training Methods2026.0671.8884.41——
GiGPOBackbone=Qwen2.5-7B, Protocol=Avg, Category=Training Methods2026.0671.7683.77-2.89-1.32
L1Backbone=Qwen2.5-3B, Protocol=Avg, Category=Training Methods2026.0671.6886.213.521.83
L2Backbone=Qwen2.5-3B, Protocol=Avg, Category=Training Methods2026.0670.8686.492.72.11
L2Backbone=Qwen2.5-7B, Protocol=Avg, Category=Training Methods2026.0670.5983.45-4.06-1.64
L2Backbone=Qwen2.5-1.5B, Protocol=Avg, Category=Training Methods2026.0670.4785.673.320.4
L1Backbone=Qwen2.5-1.5B, Protocol=Best, Category=Training Methods2026.0668.7584.51-4.69-5.31
Meta PromptBackbone=Qwen2.5-3B, Protocol=Avg, Category=Training Methods2026.0668.1684.38——
GiGPOBackbone=Qwen2.5-1.5B, Protocol=Best, Category=Training Methods2026.0667.9781.35-5.47-8.47
Meta PromptBackbone=Qwen2.5-1.5B, Protocol=Avg, Category=Training Methods2026.0667.1585.27——
GRPOBackbone=Qwen2.5-1.5B, Protocol=Best, Category=Training Methods2026.0666.4181.22-7.03-8.6
GRPOBackbone=Qwen2.5-7B, Protocol=Avg, Category=Training Methods2026.0664.0276.46-10.63-8.63
GiGPOBackbone=Qwen2.5-3B, Protocol=Avg, Category=Training Methods2026.0663.5781.49-4.59-2.89
L1Backbone=Qwen2.5-1.5B, Protocol=Avg, Category=Training Methods2026.0663.282.17-3.95-3.1
GRPOBackbone=Qwen2.5-1.5B, Protocol=Avg, Category=Training Methods2026.0662.7777.26-4.38-8.01
GiGPOBackbone=Qwen2.5-1.5B, Protocol=Avg, Category=Training Methods2026.0659.8477.68-7.31-7.59
GRPOBackbone=Qwen2.5-3B, Protocol=Best, Category=Training Methods2026.065071.02-21.88-13.39
GRPOBackbone=Qwen2.5-3B, Protocol=Avg, Category=Training Methods2026.0644.5363.9-23.63-20.48
Retro ProgressBackbone=Qwen2.5-7B, Category=Prompting Methods2026.0623.851.59.70.7
Meta Prompt for ProgressBackbone=Qwen2.5-3B, Category=Prompting Methods2026.0616.441.53.95
Random ProgressBackbone=Qwen2.5-3B, Category=Prompting Methods2026.0616.452.43.915.9
Random Retro ProgressBackbone=Qwen2.5-3B, Category=Prompting Methods2026.061648.43.511.9
Base PromptBackbone=Qwen2.5-7B, Category=Prompting Methods2026.0614.150.8——
Base PromptBackbone=Qwen2.5-3B, Category=Prompting Methods2026.0612.536.5——
Random Retro ProgressBackbone=Qwen2.5-7B, Category=Prompting Methods2026.0612.546.5-1.6-4.3
Meta Prompt for ProgressBackbone=Qwen2.5-7B, Category=Prompting Methods2026.0610.941.4-3.2-9.4
Retro ProgressBackbone=Qwen2.5-3B, Category=Prompting Methods2026.0610.644.5-1.98
Random ProgressBackbone=Qwen2.5-7B, Category=Prompting Methods2026.0610.244.4-3.9-6.4
Retro ProgressBackbone=Qwen2.5-1.5B, Category=Prompting Methods2026.065.120.74.715.7
Random Retro ProgressBackbone=Qwen2.5-1.5B, Category=Prompting Methods2026.06218.51.613.5
Meta Prompt for ProgressBackbone=Qwen2.5-1.5B, Category=Prompting Methods2026.061.23.80.8-1.2
Random ProgressBackbone=Qwen2.5-1.5B, Category=Prompting Methods2026.060.813.20.48.2
Base PromptBackbone=Qwen2.5-1.5B, Category=Prompting Methods2026.060.45——