Toolathlon Verified Pass@3
Fraction of Toolathlon Verified tasks solved in at least one of three trials.
4 models
Higher is better.
- 1Claude Opus 5Anthropic87.0
- 2Claude Sonnet 5.5Anthropic85.2
- 3Claude Opus 5.5Anthropic82.4
- 4Claude Fable 5.1Anthropic81.5
About this benchmark
- Metric
- Pass@3
- Tasks
- 108 verified real-world tool-use tasks
- Authors
- Anthropic (2026)
Scores are as reported by model developers or evaluators and may use different settings (for example reasoning effort or tools). Compare with care.