An
claude-sonnet-4-6 Benchmark & Insights
Anthropic Claude API
Updated Jun 6, 2026 All models
Sample size
79 runs
in window
Accuracy
85.2%
consensus match · 81d
Confidence
89%
over 81 runs
Window end
Jun 6, 2026
most recent run
Input price
$3.00/MTok
prompt tokens
Output price
$15.00/MTok
completion tokens
Model insights
- 01 The odd Claude out: all 12 divergences overcall, including two "high"/"unsafe" calls on consensus "safe" days.
- 02 It is more alarmist than both the cheaper Haiku and the pricier Opus, leaving it without a niche in its own family.