# tau2-bench

tau2-bench is an AI benchmark; this page records what it measures and traces its data, grader, and scores to primary sources.

Measures conversational agents when both the user and agent can change shared task state.

## Details

- Address: Online, Global
- Website: [github.com/sierra-research/tau2-bench](<https://github.com/sierra-research/tau2-bench>)
- Tags: Helpful, Science
- Category: Tools
- Status: Active
- Data licence: MIT
- Data access: open
- Grader: simulation
- Organization: Sierra Research
- Released: 2025-06
- Paper: [arxiv.org/abs/2506.07982](<https://arxiv.org/abs/2506.07982>)
- Leaderboard: [taubench.com/](<https://taubench.com/>)
- Repository: [github.com/sierra-research/tau2-bench](<https://github.com/sierra-research/tau2-bench>)
- Top-score provenance: unknown
- ABI family: No

## Questions

### Where is tau2-bench?

tau2-bench is at Online, Global.

### How do I contact tau2-bench?

To contact tau2-bench, visit https://github.com/sierra-research/tau2-bench.

## Source

From [Benchmark primary sources](https://aibenchmarkindex.com/about), license CC0-1.0, checked 2026-10-15.

Last updated October 15, 2026.

## Links

- This benchmark on the [web page](<https://aibenchmarkindex.com/global/online/tau2-bench>)
- [Tools in Online](<https://aibenchmarkindex.com/global/online/category/tools.md>)
- [More benchmarks in Online](<https://aibenchmarkindex.com/global/online.md>)

> **Check before you compare.** Benchmark methods and leaderboards change. Follow the cited primary source before using a result in a decision.