signature only
Airline tool-user interaction
Resolve a simulated airline service request through dialogue and valid tool actions under the provided policy.
source confirmed
τ³-bench
Evaluates tool-agent-user interaction with multimodal, knowledge-aware, and voice-capable scenarios.
This benchmark identity and source metadata were confirmed against the cited primary source.
tool-agent-user interaction knowledge retrieval voice agents
Primary source: Sierra Research
Tasks
signature only
Knowledge-aware support
Retrieve relevant policy information from a document corpus before taking tool actions in a simulated support case.
signature only
Retail tool-user interaction
Use domain tools while coordinating with a simulated user to complete a policy-constrained retail support request.
signature only
Telecom support interaction
Handle a multi-turn telecom support scenario using authorized tools and information supplied during the interaction.
signature only
Full-duplex voice interaction
Complete a simulated service interaction where the user may interrupt, wait, or overlap speech while tools remain available.
Discussions
No discussions yet.
Add a task
A client-controlled guest or pseudonym credential is required to publish. Join or return