reasoning benchmark
TAU-bench Airline Leaderboard
Part of τ-bench (TAU-bench), a benchmark for Tool-Agent-User interaction in real-world domains. The airline domain evaluates language agents' ability to interact with users through dynamic conversations while following domain-specific rules and using API tools. Agents must handle airline-related tasks and policies reliably.
Updated Aug 17, 2026
Models23
Model coverage23
MetricScore
EvidenceB
TAU-bench Airline Ranking
Higher score ranks better on this benchmark.
Sort by
Rank | Model | Score | Percentile | Participants | Evidence | Evaluated |
|---|
| Rank01 | ModelAN | Score70.0% | Percentile100.0% | Participants23 | EvidenceC | Evaluated |
| Rank02 | ModelMI | Score62.0% | Percentile95.5% | Participants23 | EvidenceC | Evaluated |
| Rank03 | ModelZA | Score60.8% | Percentile90.9% | Participants23 | EvidenceC | Evaluated |
| Rank04 | ModelZA | Score60.4% | Percentile86.4% | Participants23 | EvidenceC | Evaluated |
| Rank05 | ModelAN | Score60.0% | Percentile81.8% | Participants23 | EvidenceC | Evaluated |
| Rank06 | ModelMI | Score60.0% | Percentile77.3% | Participants23 | EvidenceC | Evaluated |
| Rank07 | ModelAC | Score60.0% | Percentile72.7% | Participants23 | EvidenceC | Evaluated |
| Rank08 | ModelAN | Score59.6% | Percentile68.2% | Participants23 | EvidenceC | Evaluated |
| Rank09 | ModelAN | Score58.4% | Percentile63.6% | Participants23 | EvidenceC | Evaluated |
| Rank10 | ModelAN | Score56.0% | Percentile59.1% | Participants23 | EvidenceC | Evaluated |
| Rank11 | ModelOP | Score50.0% | Percentile54.5% | Participants23 | EvidenceC | Evaluated |
| Rank12 | ModelOP | Score50.0% | Percentile50.0% | Participants23 | EvidenceC | Evaluated |
| Rank13 | ModelOP | Score49.4% | Percentile45.5% | Participants23 | EvidenceC | Evaluated |
| Rank14 | ModelOP | Score49.2% | Percentile40.9% | Participants23 | EvidenceC | Evaluated |
| Rank15 | ModelAC | Score49.0% | Percentile36.4% | Participants23 | EvidenceC | Evaluated |
| Rank16 | ModelAN | Score46.0% | Percentile31.8% | Participants23 | EvidenceC | Evaluated |
| Rank17 | ModelAC | Score46.0% | Percentile27.3% | Participants23 | EvidenceC | Evaluated |
| Rank18 | ModelAC | Score44.0% | Percentile22.7% | Participants23 | EvidenceC | Evaluated |
| Rank19 | ModelOP | Score42.8% | Percentile18.2% | Participants23 | EvidenceC | Evaluated |
| Rank20 | ModelOP | Score36.0% | Percentile13.6% | Participants23 | EvidenceC | Evaluated |
| Rank21 | ModelOP | Score32.4% | Percentile9.1% | Participants23 | EvidenceC | Evaluated |
| Rank22 | ModelAN | Score22.8% | Percentile4.5% | Participants23 | EvidenceC | Evaluated |
| Rank23 | ModelOP | Score14.0% | Percentile0.0% | Participants23 | EvidenceC | Evaluated |
TAU-bench Airline Highlights
The leading models and scores on this benchmark.
TAU-bench Airline Score Distribution
A closer view of the leading scores on this benchmark.
The Top AI Models for TAU-bench Airline
The first five results on this benchmark, with official price and output speed added where the model identity can be matched.
Ranking basisThis tau-bench airline AI model leaderboard uses descending score in the benchmark's original unit. The leaderboard ranking keeps matched price and speed data separate from benchmark evidence.
- 01ANClaude Sonnet 4.5Anthropic
- Score
- 70.0%
- Price
- $3.0 input / $15 output per 1M tokens
- Speed
- Up to 42 tok/s via Anthropic
Strengths
- Ranks #1 of 23 compared models
- 100th percentile on this benchmark
- C evidence result
Considerations
- This result measures TAU-bench Airline, not total model capability
- 02MIMiniMax M1 80KMiniMax
- Score
- 62.0%
Strengths
- Ranks #2 of 23 compared models
- 95th percentile on this benchmark
- C evidence result
Considerations
- This result measures TAU-bench Airline, not total model capability
- 03ZAGLM-4.5-AirZhipu AI
- Score
- 60.8%
- Price
- $0.20 input / $1.1 output per 1M tokens
Strengths
- Ranks #3 of 23 compared models
- 91th percentile on this benchmark
- C evidence result
Considerations
- This result measures TAU-bench Airline, not total model capability
- 04ZAGLM-4.5Zhipu AI
- Score
- 60.4%
- Price
- $0.60 input / $2.2 output per 1M tokens
Strengths
- Ranks #4 of 23 compared models
- 86th percentile on this benchmark
- C evidence result
Considerations
- This result measures TAU-bench Airline, not total model capability
- 05ANClaude Sonnet 4Anthropic
- Score
- 60.0%
- Speed
- Up to 101 tok/s via Bedrock
Strengths
- Ranks #5 of 23 compared models
- 82th percentile on this benchmark
- C evidence result
Considerations
- This result measures TAU-bench Airline, not total model capability
Selection summary
Best AI Models for TAU-bench Airline
Claude Sonnet 4.5 currently leads TAU-bench Airline with 70.0%. It is the top model on this specific benchmark, while the best LLM for the broader task should also be checked against other benchmarks, price and runtime.
Use this leaderboard with the supporting benchmark results and coverage details above. A leaderboard position summarizes the selected ranking signal; it does not replace workload-specific testing.
What is TAU-bench Airline?
What TAU-bench Airline measures and how its scores work.
Part of τ-bench (TAU-bench), a benchmark for Tool-Agent-User interaction in real-world domains. The airline domain evaluates language agents' ability to interact with users through dynamic conversations while following domain-specific rules and using API tools. Agents must handle airline-related tasks and policies reliably.
Scores are shown in ratio. This benchmark is not independently verified and has an evidence level of B.
- Family
- TAU-bench Airline
- Modality
- text
- Primary category
- reasoning
- Score direction
- higher
- LLMBoard eligible
- Yes
- Evaluation key
- tau-bench-airline|llm-stats-current
Benchmark scores retain their original unit. Overall score eligibility is shown separately.
FAQ
Common questions about TAU-bench Airline.
Which model scores highest on TAU-bench Airline?
Claude Sonnet 4.5 is currently ranked first with 70.0%.
What does TAU-bench Airline measure?
Part of τ-bench (TAU-bench), a benchmark for Tool-Agent-User interaction in real-world domains. The airline domain evaluates language agents' ability to interact with users through dynamic conversations while following domain-specific rules and using API tools. Agents must handle airline-related tasks and policies reliably.
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
23 model results are currently shown.
Does this benchmark affect the overall score?
Yes. This benchmark can contribute to the current LLMBoard capability score.