Skip to content

Benchmark

Toolathlon Verified · Pass@3

Scores are percentages. Higher is better.

Models measured
6
Scores recorded
6
Sources
1
IQ weight
No weight published

Score distribution

84.7%
Mean
85.2%
Median
2.3%
Std dev
82.4%
10th percentile
88%
90th percentile
81.5%
Min
81.5%88%
The band covers the middle 80% of scores. The line marks the median.

Leaderboard

#ModelScore
1Claude Opus 4.8internal pinned harness · adaptive thinking · max · 108 tasks · 3 trials· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.88%
2Claude Opus 5internal pinned harness · adaptive thinking · max · 108 tasks · 3 trials· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.87%
3Claude Sonnet 5.5internal pinned harness · adaptive thinking · max · 108 tasks · 3 trials· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.85.2%
4Claude Sonnet 5internal pinned harness · adaptive thinking · max · 108 tasks · 3 trials· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.84.3%
5Claude Opus 5.5internal pinned harness · adaptive thinking · max · 108 tasks · 3 trials· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.82.4%
6Claude Fable 5.1internal pinned harness · adaptive thinking · max · 108 tasks · 3 trials· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.81.5%

Data credits: BenchBoard by SYuan03 (github.com/SYuan03/benchboard). MIT covers its code and data organization; no results licence is inferred. Numeric relays retain the original stating-source links.