Claude Fable 5.1's language is less "load-bearing" than its predecessor's
decoder ·
Arena.ai analyzed tens of thousands of high-reasoning Text Arena outputs to compare how Anthropic's Claude writing changed from Fable 5 to Fable 5.1. The newer model uses fewer agreement openers, fewer em dashes, and less wording like "honestly" and "frankly," while its answers have grown longer. Specifically, Fable 5.1 uses 11.0 em dashes per 1,000 words versus Fable 5's 16.2, but relies more on semicolons, rising to 6.09 from 3.73. Its median answer length is 414 words, 30 percent more than Fable 5's 319 words but 21 percent shorter than Opus 5's 525 words. Fable 5.1 also shows a 20 percent drop in stock phrases such as "load-bearing," a 36 percent decline in hedges like "perhaps" and "arguably," and praise or validation in 1.98 percent of responses, down from 3.17 percent. Long content words fall from 42.6 percent to 38.6 percent, abstract nouns drop from 4.39 to 3.28 per 100 words, and sentence length is 12.54 words, slightly above Opus 5's 12.20.
decoder