Benchmark · Agent tasks
τ²-bench
A model working as a customer-service agent under a written policy, with tools and a simulated customer, in one domain at a time (airline, retail, telecom, banking), in text or by voice. Each domain is scored on its own.
- Score
- % of tasks passed on the first try (pass^1), higher is better
- Run by
- Sierra
- Status
- Current, newest result 29 October 2026
- Updated
- as submissions land
- Data
- the published file, MIT, read 20 November 2026
- Method
- taubench.com/
Results: text: banking knowledge
4 models we list, each at its best setting, best first.
| # | Model | Score | Range | Who ran it | Setting | Date | |
|---|---|---|---|---|---|---|---|
| 1 | Kimi K3 B | 37.1% | Independent | effort max, retrieval alltools, submission type standard, tau2 version 1.0.1, user simulator gpt-5.2 | 2026-08-04 | source | |
| 2 | Gemini 3.1 Pro (preview) AA | 26.0% | Independent | effort high, retrieval alltools, submission type standard, tau2 version 1.0.1, user simulator gpt-5.2 | 2026-05-05 | source | |
| 3 | Inkling B | 25.0% | Independent | effort max, retrieval alltools, submission type standard, tau2 version 1.0.1, user simulator gpt-5.2 | 2026-08-04 | source | |
| 4 | Qwen3.5-397B-A17B B | 9.8% | Independent | effort enabled, retrieval text-emb-3-large, submission type standard, tau2 version 1.0.1, user simulator gpt-5.2 | 2026-03-02 | source |
Results: voice: banking knowledge
1 model we list, each at its best setting, best first.
Results: voice-v2-entity
1 model we list, each at its best setting, best first.
Results: voice-v2-banking
1 model we list, each at its best setting, best first.
Results: voice-v2
1 model we list, each at its best setting, best first.
Results: text: telecom (legacy)
1 model we list, each at its best setting, best first.
| # | Model | Score | Range | Who ran it | Setting | Date | |
|---|---|---|---|---|---|---|---|
| 1 | DeepSeek-V3.2 C | 96.2% | Submitted by the maker | tau2 version v0.1.3, user simulator DeepSeek-V3.2 | 2025-12-01 | source |
Results: text: telecom
1 model we list, each at its best setting, best first.
| # | Model | Score | Range | Who ran it | Setting | Date | |
|---|---|---|---|---|---|---|---|
| 1 | Qwen3.5-397B-A17B B | 97.8% | Independent | effort enabled, submission type standard, tau2 version 1.0.1, user simulator gpt-5.2 | 2026-03-02 | source |
Results: text: retail (legacy)
1 model we list, each at its best setting, best first.
| # | Model | Score | Range | Who ran it | Setting | Date | |
|---|---|---|---|---|---|---|---|
| 1 | DeepSeek-V3.2 C | 81.1% | Submitted by the maker | tau2 version v0.1.3, user simulator DeepSeek-V3.2 | 2025-12-01 | source |
Results: text: retail
1 model we list, each at its best setting, best first.
| # | Model | Score | Range | Who ran it | Setting | Date | |
|---|---|---|---|---|---|---|---|
| 1 | Qwen3.5-397B-A17B B | 84.4% | Independent | effort enabled, submission type standard, tau2 version 1.0.1, user simulator gpt-5.2 | 2026-03-02 | source |
Results: text: airline (legacy)
1 model we list, each at its best setting, best first.
| # | Model | Score | Range | Who ran it | Setting | Date | |
|---|---|---|---|---|---|---|---|
| 1 | DeepSeek-V3.2 C | 63.8% | Submitted by the maker | tau2 version v0.1.3, user simulator DeepSeek-V3.2 | 2025-12-01 | source |
Results: text: airline
1 model we list, each at its best setting, best first.
| # | Model | Score | Range | Who ran it | Setting | Date | |
|---|---|---|---|---|---|---|---|
| 1 | Qwen3.5-397B-A17B B | 81.5% | Independent | effort enabled, submission type standard, tau2 version 1.0.1, user simulator gpt-5.2 | 2026-03-02 | source |
A rank is among the models we list with a result, not a place on the whole leaderboard, and a model offered by several hosts is shown once, on its own listing. These results don't move our grades. Data: τ²-bench, Sierra Research, MIT, MIT.
For companies
Do agents find, use and choose your tools?
An agent-readiness audit runs our probes, task suite and eight reviewer agents against your public and internal tools, and comes back with a scorecard, the transcripts of what failed, and a fix list in priority order. From $2,500, re-run included. We never take payment to move a rank. We do help companies earn one.