Benchmark
Tau2 Airline
reasoning
communication
text
TAU2 airline domain benchmark for evaluating conversational agents in dual-control environments where both AI agents and users interact with tools in airline customer service scenarios. Tests agent coordination, communication, and ability to guide user actions in tasks like flight booking, modifications, cancellations, and refunds.
Top 10
榜单概览
数据说明
分数为 0-100 百分制综合成绩（Elo 榜为相对评分，仅供同榜内比较）。评测分数属事实数据，整理自公开评测体系与厂商公开资料，部分分数为厂商自报口径，不同榜单的评测方法与数据集可能存在差异，请结合多个榜单交叉参考。
Ranking
模型排名
| 名次 | 模型 | 机构 | 分数 | 来源 |
|---|---|---|---|---|
| 1 | Claude Haiku 4.5 | Anthropic | 63.6 | 来源 |
| 2 | GPT-5 | OpenAI | 62.6 | 来源 |
| 3 | Qwen3-Next-80B-A3B-Thinking | Alibaba Cloud / Qwen Team | 60.5 | 来源 |
| 4 | Qwen3-235B-A22B-Thinking-2507 | Alibaba Cloud / Qwen Team | 58.0 | 来源 |
| 5 | Kimi K2-Instruct-0905 | Moonshot AI | 56.5 | 来源 |
| 6 | Qwen3-Next-80B-A3B-Instruct | Alibaba Cloud / Qwen Team | 45.5 | 来源 |