Simon Willison·· 2026-09-29
Claude Sonnet 5.5
Claude Sonnet 5.5
AI summary
Claude Sonnet 5.5 New Sonnet model from Anthropic today. They say it "runs 30%+ faster, and costs up to 30% less for most work" - it's priced the same as Sonnet 5 but appears to beat it on every benchmark, and should be cheaper to run as well. Here are some pelicans riding bicycles .
Selection record
Threshold 76Media and individualsFirst 84Second 78
AdmittedSum of both 162 ≥ twice the threshold 152
- Source tier
- Media and individuals; this tier's threshold is 76
- Pre-filter
- passed：介绍Anthropic新模型性能与定价
- Why it was chosen
- The author tests Sonnet 5.5 against Opus 5.5 for coding and notes that it is now the claude.ai free-tier model.
- Same event
A model scores each item twice, independently, against one written standard, out of 100. An item is admitted only when the two scores add up to twice the threshold. The threshold is set per source tier.
Source: Simon Willison · simonwillison.net