# τ²-bench Airline leaderboard

Agentic benchmark

> Every figure below is reproduced as its upstream source published it: nothing is modelled, estimated, interpolated or converted. `Unavailable` means no source published the value — it is never a zero. Each value carries its evidence state and the date it was observed.

Page: https://modelchoice.dev/leaderboards/benchmarks/agentic-tau2-airline

τ²-Bench Airline Domain. Every model the catalog carries a published τ²-bench Airline value for, ranked by that value.

## Benchmark record

| Field | Value |
| --- | --- |
| Category | Agentic |
| Measure | Domain success under a published trial policy |
| Tasks | Airline customer-service tasks |
| Difficulty | Policy-constrained airline support workflows |
| Published by | [τ²-Bench: Evaluating Conversational Agents in a Dual-Control Environment](https://arxiv.org/abs/2506.07982) |

τ²-bench Airline tests conversational agents on airline customer-service tasks governed by domain policy and database-changing tools.

## τ²-bench Airline ranking

5 models with a published τ²-bench Airline value, ordered by that value, highest first. Models the source has not scored on this benchmark are not listed — they are unmeasured, not last.

| Rank | Model | Provider | Domain success under a published trial policy | Evidence |
| --- | --- | --- | --- | --- |
| 1 | [Trinity-Large-Thinking](https://modelchoice.dev/models/trinity-large-thinking) | Arcee AI | 88.0 | Observed 2026-11-21 · source benchlm:models |
| 2 | [Nemotron 3.5 Super VL 120B A12B](https://modelchoice.dev/models/nemotron-3-5-super-vl-120b-a12b) | NVIDIA | 76.5 | Observed 2026-11-21 · source benchlm:models |
| 3 | [Nemotron 3.5 Super VL 120B A12B NVFP4](https://modelchoice.dev/models/nemotron-3-5-super-vl-120b-a12b-nvfp4) | NVIDIA | 75.8 | Observed 2026-11-21 · source benchlm:models |
| 4 | [Ember-1](https://modelchoice.dev/models/ember-1) | Fireworks | 66.0 | Observed 2026-11-21 · source benchlm:models |
| 5 | [ZAYA1-74B-Preview](https://modelchoice.dev/models/zaya1-74b-preview) | Zyphra | 56.1 | Observed 2026-11-21 · source benchlm:models |

Rows are ordered by the value the source published, highest first. The source does not state whether a higher value is the better result, so this document does not either: for a benchmark that measures a rate of failure, read the table from the bottom.
