Anthropic's Claude Opus 5.5 appears to be changing how it writes. New analysis shows fewer obvious AI writing patterns, shorter sentences, and simpler wording compared with Opus 5.
Claude Opus 5.5 stands out as one of the best models for coding. It also appears to be a bit better at writing as Anthropic shifts how the AI communicates.
Data from Arena, an AI benchmarking tool, reveals that Opus 5.5 features fewer obvious AI writing patterns. Users are less likely to create AI slop content with this model.
Read nextAnthropic Offers Up to $250 in Free Claude Code Cloud Session CreditsArena analyzed high-reasoning Text Arena responses from August and September 2026. The evaluation observed that 10 of 12 writing measures moved in what Arena considers a better direction.
Arena data confirms that Opus has almost stopped using em dashes. Em dashes previously served as one of the biggest signs of AI-generated content.
Opus 5 used 15.2 em dashes per 1,000 words. Opus 5.5 dropped that figure to just 0.8, marking a reduction of roughly 95%.
Semicolon usage also fell sharply from 6.10 to 1.64 per 1,000 words. The newer model writes shorter sentences, averaging 10.03 words compared with 12.14 for Opus 5.
One obvious tradeoff accompanies these updates. Opus 5.5 is more verbose than its predecessor.
Average answers increased from 453 to 481 words. This makes it the longest-writing Opus model included in the comparison.
AI models have recently focused mostly on coding capabilities. Watching Anthropic change how Claude writes offers an interesting shift, resulting in fewer em dashes across the internet.
