τ-Bench

Benchmarks

Task Name	Dataset Name	SOTA Result
Agentic Reasoning	τ-Bench	Score62.58	100
Tool-use Agent Performance	τ²-bench	ASR72.4	50
Tool-use	τ-Bench	Average Pass@167.42	38
Long-context Reasoning	∞ Bench	Accuracy90.39	32
Agent Task Completion	τ-BENCH (test)	Average Task Reward0.791	27
User Simulator Goal Alignment	τ-Bench Retail (test)	User Profile Success Rate94.5	19
Conversational Tool-use	τ²-Bench	Airline Success Rate75.5	18
Behavioral Similarity Analysis	τ-Bench and τ2-Bench (test)	GED Score82.6	18
Agentic Tool Use	τ²-Bench Telecom	Accuracy100	18
Question Answering	∞-Bench Longbook QA English (test)	F1 Score11.2	18
User simulator goal alignment	τ-Bench Retail	User Profile Adherence94.5	14
User simulator goal alignment	τ-Bench Airline	User Profile Alignment (Prof.)98.7	14
Tool Use Reasoning	τ-Bench	Avg Accuracy63.9	14
Tool Use	τ-Bench (TauB) V2	Accuracy91.6	13
Long-context language tasks (MC, QA, Sum)	∞Bench	MC Accuracy78.6	13
Long-context Question Answering	∞Bench	Accuracy78.46	13
Story Question Answering	∞Bench En.MC	Accuracy90	12
Customer Support Interaction	τ-Bench Telecom Verified (test)	Pass@194	11
Customer Support Interaction	τ-Bench Retail Verified (test)	Pass Rate92	11
Customer Support Interaction	τ-Bench Airline Verified (test)	Pass@182	11
Long-Text Understanding	∞BENCH (test)	Overall Accuracy85.6	10
End-to-end task completion	τ-Bench Retail, N=5	Task Completion Rate0.111	8
Tool Use	τ²-Bench (out-of-distribution)	Retail Score54.9	8
Question Answering	∞Bench Zh.QA	F1 Score49.1	7
Question Answering	∞Bench En.QA	F1 Score42.1	7

Showing 25 of 51 rows