Anthropic has launched Claude Fable 5.2 and Claude Mythos 5.2, two upgraded models for complex technical work. The models generate outputs up to 30% faster than their predecessors. Pricing is $8 per million input tokens and $40 per million output tokens.
Performance and cost
On FrontierCode 2.0, Fable 5.2 at High effort scored 45.3%, close to Opus 5.5 at 45.5% for about half the cost per task. At Xhigh effort, it reached 49.1%, the highest released result tested. The model also matched Opus 5.5's top score on Humanity's Last Exam at Low effort for less than one-tenth of the cost.
Safety and access
Fable 5.2 includes cybersecurity and biology safeguards, while Mythos 5.2 is available to vetted organizations. Anthropic says Fable 5.2 produces about 55% fewer unnecessary blocks than Fable 5.1. Access for Mythos runs through the Cyber Verification Program and Life Sciences Verification Program.
Scientific and engineering results
Anthropic says prerelease testing helped complete two open problems in black hole physics. A team of 1,000 Fable 5.2 agents fully formalized Fermat's Last Theorem in 30 hours. The model predicted whole capsid protein structures in 28 of 34 attempts.
Claude Fable 5.2 is available now on Amazon Web Services, Google Cloud, and Microsoft Azure. Fast mode is offered on Claude Code, the Claude Platform, Google Cloud, and Microsoft Foundry. Paid plan subscribers will receive extra Fable usage this week.