Anthropic's Fable 5.2: 1,000+ agents wrote a bootable OS in eight days, yet the company says "start with Opus"

Developer Theo posted on Nov. 9 (US time): "It's been under an hour and I'm already struggling with 'should I use Opus 5.5 or Fable 5.2?'" (about 2,400 likes). Anthropic released Claude Fable 5.2 the same day as its top model, yet told users to start most coding and knowledge work on Opus 5.5.

In its blog post Anthropic said Fable 5.2 generates output up to 30% faster than 5.1 and costs $8 per million input tokens and $40 per million output tokens, down from $10 and $50. It also shared test cases: one engineer started from an empty folder and ran more than 1,000 subagents for eight days, producing an operating system that boots on a laptop. From a one-paragraph prompt the model designed three working electronic devices, with boards ready for manufacture in under a day. It formalized the proof of Fermat's Last Theorem in 30 hours, and Anthropic said more than 90% of changes merged into its core codebase are now written by Claude.

Early reaction is mostly positive. Wharton professor Ethan Mollick, who had early access, said the model "crosses the line into being a sort of Jagged AGI" (about 1,900 likes). Cursor said Fable 5.2 scores 60% on its CursorBench at max effort and 56.9% at default effort, up from 49.2% for Fable 5.1. Anthropic's Felix Rieseberg wrote that Fable beats Opus on big, long tasks, but "I often still prefer Opus 5.5 for everyday work. Fable's got a big brain and thinks a little slower."

Cost complaints came fast. When an Anthropic engineer posted a game made with Fable, one reply said it is easy to go AAA "when you're not on the same weekly limit as the rest of us." Anthropic is giving subscribers Fable usage resets that grow with how long they have subscribed: one for Pro, up to three for Max, Ultra and Team Premium, Boris Cherny said (about 10,000 likes). Benchmarks and cost comparisons are in the feature below.