List your company

Customer Support Interaction on τ-Bench Retail Verified (test)

92Pass RateGPT-5.2
45.257.3569.581.65May 27, 2026

Evaluation results

MethodLinks
GPT-5.2User Simulator=Gemini-3-flash2026.0592
Gemini-3-flashUser Simulator=Gemini-3-flash2026.0588
Gemini-3-flashUser Simulator=GPT-5.22026.0587
Gemini-2.5-flashUser Simulator=Gemini-3-flash2026.0579
GPT-5.2User Simulator=GPT-5.22026.0569
claude-sonnet-4.6User Simulator=Gemini-3-flash2026.0567
deepseek-3.1User Simulator=GPT-5.22026.0567
Gemini-2.5-flashUser Simulator=GPT-5.22026.0560
Qwen-32BUser Simulator=GPT-5.22026.0556
Qwen-32BUser Simulator=Gemini-3-flash2026.0548
deepseek-3.1User Simulator=Gemini-3-flash2026.0547