Tau2-Bench Verified: leaderboard
An agent model serves simulated customers (played by gpt-5.1) in airline, retail and telecom settings, following a written policy and calling tools; the score is the unweighted mean of the three domain pass rates on the 278 base tasks (50 airline, 114 retail, 114 telecom); the README says its figures are averaged over 5 runs. Unlike tau2-bench, the tasks whose expected actions broke the policy or the database were corrected. Amazon's own Nova 2 rows were added in a later commit that does not state how they were run.
Metric: Mean of the airline, retail and telecom pass^1 (%). Source: github.com. Saturation forecast: Around February 2027. 14 models tracked. Updated 17 Nov 2026.
Top models
| # | Model | Setting | Score | Overall rank |
|---|---|---|---|---|
| 1 | Grok 4.1 Fast | Reasoning | 82.71 | #429 |
| 2 | Claude Opus 4.5 | 81.99 | #161 | |
| 3 | GPT-5.2 | xHigh | 80.3 | #185 |
| 4 | GPT-5 | Medium | 79.92 | #177 |
| 5 | Gemini 3 Pro | 79.39 | #145 | |
| 6 | Nova 2 Pro | 78.53 | #516 | |
| 7 | GPT-5.1 | High | 76.82 | #235 |
| 8 | Nova 2 Omni | 75.7 | #472 | |
| 9 | Claude Sonnet 4.5 | 73.32 | #258 | |
| 10 | Nova 2.0 Lite | 72.43 | #1125 | |
| 11 | GPT-5 Mini | Medium | 69.83 | #359 |
| 12 | Gemini 2.5 Pro | 56.38 | #274 | |
| 13 | Claude Haiku 4.5 | 53.44 | #580 | |
| 14 | Gemini 2.5 Flash | 41.57 | #467 |
No result here: #1 Claude Fable 5.2, #2 GPT-6.2 Astra, #3 Gemini 4 Carbon.
Interactive version: theaggregate.ai/benchmark?slug=tau2-bench-verified · How It Works · Data refreshed daily, snapshot 2026-11-17.