List your company

Trajectory Search on WebArena Lite 1.0 (test)

44.44Shopping SuccessWebArbiter-7B
20.83226.96133.0939.219Jan 29, 2026

Evaluation results

MethodLinks
WebArbiter-7BPolicy=GPT-4o2026.0144.4442.8652.6356.6749.1514.11
WebArbiter-7BPolicy=GPT-4o-mini2026.0137.7842.8636.8446.6741.0419.13
WebShepherd-8BPolicy=GPT-4o2026.0130.4342.8647.6246.8841.956.91
GPT-4o-miniPolicy=GPT-4o2026.0126.6737.1442.114036.481.44
WebShepherd-8BPolicy=GPT-4o-mini2026.0126.0945.7123.8140.6234.069.55
GPT-4o-miniPolicy=GPT-4o-mini2026.0124.4422.8626.3233.3326.742.23
w/o Trajectory SearchPolicy=GPT-4o2026.0123.9131.4328.5756.2535.04—
w/o Trajectory SearchPolicy=GPT-4o-mini2026.0121.7422.8619.0534.3824.51—