Skip to content

Benchmark

tau-bench original airline pass^1

Agentic benchmarks measure how well models carry out tasks that take several steps. Scores are percentages. Higher is better.

Models measured
3
Scores recorded
3
Sources
1
IQ weight
No weight published

Score distribution

58.5%
Mean
59.6%
Median
1.8%
Std dev
56%
10th percentile
60%
90th percentile
56%
Min
56%60%
The band covers the middle 80% of scores. The line marks the median.

Leaderboard

#ModelScore
1Claude Sonnet 4submission-claude-sonnet-4_anthropic_2025-05-22· Sierra tau-bench Leaderboard data docs60%
2Claude Opus 4submission-claude-opus-4_anthropic_2025-05-22· Sierra tau-bench Leaderboard data docs59.6%
3Claude Opus 4.1submission-claude-opus-4-1_anthropic_2025-01-15· Sierra tau-bench Leaderboard data docs56%

Data credits: Sierra Research tau2-bench leaderboard submissions (github.com/sierra-research/tau2-bench), MIT (licence)