Claude Fable 5.1 cuts agreement openers 58% and em dashes 32%, Arena analysis finds
rohanpaul_ai · x · 2026-09-10
An analysis of tens of thousands of high-reasoning Text Arena outputs shows notable style shifts from Fable 5 to Fable 5.1:
- Agreement openers like "yes" and "exactly" fell 58%, from 2.35% to 0.99% of responses
- Honesty wording like "honestly" and "frankly" dropped 45% per 1,000 words
- Em dashes fell 32% per 1,000 words; praise dropped from 3.17% to 1.98%
- Semicolons rose 63% per 1,000 words, and answers have grown longer
Overall, Anthropic appears to be deliberately scrubbing the trademark "AI-isms" users have long complained about.
More from Models
- Codex CLI 0.154.0 ships GPT-6-Astra, experimental worktree support — github-actions[bot] · 2026-09-10
- OpenAI's new model, in training since Aug 28, reportedly beat GPT-6-Astra in just one week — alexcovo_eth · 2026-09-10
- Qwen3.8-2.4T-A95B open weights land on AWS: single 8×B300 node with vLLM — AWS ML Blog · 2026-09-10
- Users say Astra's $200 sub is no longer enough: multi-project work burns through quota in days — CtrlAltDwayne · 2026-09-10
- OpenAI launches GPT-6 Astra to power ChatGPT Work with desktop app control — OpenAI · 2026-09-10
- Intelligence Index v4.3: Claude Fable 5.1, Muse Spark 1.3 and GPT-6 Astra Reset the Cost-Efficiency Frontier — ArtificialAnlys · 2026-09-10