II.
Benchmark overview
Reference · livebenchmark:tau-bench
tau-bench overview
Sierra Research's tau-bench — multi-turn agent benchmark in retail and airline customer-service domains.
Attributes
displayName
tau-bench
homepageUrl
kind
tool-use
targetsKind
AgentVersion
description
Sierra Research's tau-bench — multi-turn agent benchmark in retail and airline customer-service domains.
Outgoing edges
applies_to2
- domain:retail·DomainRetail
- domain:customer-experience·DomainCustomer Experience
covers2
- skill-area:multi-turn-tool-use·SkillAreaMulti-Turn Tool Use
- skill-area:agentic-loops·SkillAreaAgentic Loops
Incoming edges
belongs_to_benchmark1
- test-set:tau-bench-airline-retail·TestSettau-bench airline and retail domains
bounds_subject1
- scope-boundary:tau-bench.scope·ScopeBoundary