τ-bench (tau-bench)

by Sierra Research

Editor note

Simulates a customer talking to an AI agent in domains like retail and airlines. The agent must follow company policy and use tools; the final database state is checked.

What it tests: Policy-following support agents with tools, plus consistency across repeated runs.
Original source: github.com/sierra-research/tau-bench
Last updated: Repo last commit 2026-03-18
License: MIT

Facts come only from the linked source (checked 15 Oct 2026). Credit belongs to the creators.

Reviews

No reviews yet.

Reviews from practitioners who sign in with GitHub open soon.