Share your thoughts, 1 month free Claude Pro on usSee more

Web Agent Navigation on MiniWoB (full)

77.1Success Rate

Gemini 3.1 Pro

Updated 9d ago

Evaluation Results

Method	Links
Gemini 3.1 Pro 2026.04		77.1
Qwen3.5-27B 2026.04		70.9
Claude 3.5 Sonnet 2026.04		69.8
A3-Qwen3.5-9B 2026.04		69
o1-mini 2026.04		67.8
Llama 3.1 405B 2026.04		64.6
GPT-4o 2026.04		63.8
Qwen3.5-9B (base) 2026.04		63.2
Llama 3.1 70B 2026.04		57.6
GPT-4o-mini 2026.04		56.6