
Anthropic launches Claude Fable 5.2 with lower API pricing, available to all paid users
Anthropic has released Claude Fable 5.2 and Claude Mythos 5.2, introducing improvements in coding, reasoning, scientific research, cybersecurity, and biomedical work. Both models share the same architecture, although Fable has stricter cybersecurity and biology safeguards for general users. Anthropic says the models generate output up to 30% faster while reducing API costs by 20%, with input tokens priced at $8 per million and output tokens at $40 per million. An optional fast mode offers up to 2.5 times faster token generation at twice the standard API price.
On benchmarks, Fable 5.2 scored 49.1% on FrontierCode 2.0 and 74.2% on Humanity's Last Exam with tools, outperforming GPT-6 Astra's respective scores of 42.8% and 57.2%. Anthropic also demonstrated the model's capabilities through several projects, including coordinating over 1,000 AI agents to build a bootable operating system in eight days, complete with drivers, file storage, a windowing system, and an installer (yes, a friking OS... 🤷♂️). Other demonstrations included formalizing the proof of Fermat's Last Theorem in 30 hours and designing three electronic devices that worked after manufacturing without electrical modifications.
Claude Fable 5.2 is now available to all paid Claude users, with additional usage offered to eligible subscribers. Developers can also access it through Claude Code, the Claude API, AWS, Google Cloud, and Microsoft Azure using the claude-fable-5-2 model ID. Meanwhile, Claude Mythos 5.2 is available to verified cybersecurity and biomedical research organizations through Anthropic's verification programs, with enterprise access through Claude Security.





Comments
"A team of agents running an earlier version of Fable 5.2 migrated a critical production service from Python to Rust—the service is now more efficient, more reliable, and starts about 26 times faster." Anthropic, you've got any proof of it, or we can safely assume you've just made all this up? "As an experiment, one Anthropic engineer gave Fable 5.2 the task of designing a prototype operating system." Again, Anthropic, you've spent millions on it, can we we try it? Can we at least see even few lines of code? Anthropic: You're absolutely right! We're providing an awful 8s video of it to encourage you to burn trillion tokens to get your own. If you have a question about our plans and enterprise solutions, you can ask me.
It's just about a prompt slot machine, not particularly interesting. Boring, in fact!