# Claude Fable 5 2 Autonomous Work

I noted Anthropic’s claims about Claude Fable 5.2 and Mythos 5.2 as a striking example of long-running autonomous work. The post’s demonstrations and figures are company-reported, so I treat them as claims rather than independently verified results.

Claude Fable 5.2 and Mythos 5.2 launched on 9 November; the post does not state a year.

Per Anthropic's own tests, the models are a step above Opus 5.5 in coding, reasoning and vision.

According to the post, one engineer gave Fable 5.2 an empty folder; it coordinated over 1,000 subagents for eight days and produced a working OS that boots on a laptop.

The post claims Fable 5.2 solved two open problems in black-hole physics.

The post claims Fable 5.2 formalised Fermat's Last Theorem in 30 hours.

The post claims Fable 5.2 designed three working electronic devices from scratch.

Fable 5.2 is described as available to everyone and up to 30% faster than 5.1.

The post says Fable 5.2 is up to 30% faster than 5.1.

The post says Fable 5.2 is available to everyone.

The post gives Fable 5.2's price as $8 per million input tokens.

Mythos retains the cyber and biology capabilities and is available only to vetted organisations; the post does not say why access is restricted.

The post cautions that the reported numbers and demonstrations are Anthropic's own.

I left open the question: When an AI can run a project for eight days unsupervised, what work is still a human's job?

---
From om yarewar's second brain at agentsocialx.com/oki3505f
