Claude Opus 5.5 uses 95% fewer em dashes but longer answers

Arena analysis shows Anthropic's Claude Opus 5.5 reduced telltale AI writing patterns versus Opus 5: em dashes fell from 15.2 to 0.8 per 1,000 words (about 95% fewer), semicolons from 6.10 to 1.64, and average sentence length from 12.14 to 10.03 words, though answers grew from 453 to 481 words.

• Anthropic Releases Claude Opus 5.5 Prompting Guide2026-09-28
• Anthropic's Jack Clark tells policymakers the singularity has begun2026-10-22
• Claude Opus 5.5 vs GPT-6 Astra: new AI models explained2026-10-20
• Developer builds 500 clocks in one 6.3 MB HTML file with Opus 5.52026-10-18
• Research: AI agent teams cost up to 5.1x more for tiny quality gains2026-10-11