An
claude-opus-4-8 Benchmark & Insights
Anthropic Claude API
Updated Jul 31, 2026 All models
Sample size
169 runs
in window
Accuracy
94.2%
consensus match · 206d
Confidence
82%
over 207 runs
Window end
Jul 31, 2026
most recent run
Input price
$5.00/MTok
prompt tokens
Output price
$25.00/MTok
completion tokens
Model insights
- 01 The strongest model with full coverage, and its errors are one-sided: it never overrates, it only misses downward.
- 02 What matters is where — on the two "high" consensus days in July it said "medium" and flagged the day "safe", so a premium-priced model softens exactly the days you want a warning on.
Notes
Improved version of Opus 4.7. Release notes for Claude Opus 4.8