List your company

Tool-use Task Completion on Tau2-Bench Telecom subset Solo mode

43.3Success RateReAct + EvoSOP
35.08437.21739.3541.483Jul 8, 2026

Evaluation results

MethodLinks
ReAct + EvoSOPBackbone=Gemini-3-FP2026.0743.3
ReAct + EvoSOPBackbone=GPT-4o2026.0740.9
ReAct + EvoSOPBackbone=Qwen-Max2026.0740.2
ReAct + ASIBackbone=Qwen-Max2026.0739.5
ReAct + ASIBackbone=GPT-4o2026.0737.7
ReActBackbone=GPT-4o2026.0737.1
ReAct + ASIBackbone=Gemini-3-FP2026.0735.4