Anthropic launches Claude Fable 5.2 and Mythos 5.2

Anthropic released Claude Fable 5.2 and Claude Mythos 5.2 on November 9, 2026, framing both as upgrades for hard technical work in coding, reasoning and vision. Standard API pricing is $8 per million input tokens and $40 per million output tokens, down from $10 and $50 on Fable 5.1 and Mythos 5.1, according to Anthropic and the Claude Platform release notes. Anthropic also says the new models generate outputs up to 30% faster than the 5.1 versions. The Verge covered the launch the same day; the prices and speed claim above are from Anthropic’s materials.

Fable and Mythos are the same underlying model with different safeguards. Fable 5.2 is available now on Anthropic’s platforms, including Amazon Web Services, Google Cloud and Microsoft Azure, under the model ID claude-fable-5-2. Mythos 5.2 (claude-mythos-5-2) is limited to organizations verified through Anthropic’s Cyber and Life Sciences verification programs. Both have a 1M-token context window, 128K max output tokens and always-on adaptive thinking, the release notes say. An optional fast mode costs $16 per million input tokens and $80 per million output and, Anthropic says, speeds token generation by up to 2.5 times. For everyday coding and knowledge work, Anthropic still points builders to Claude Opus 5.5 first, then up to Fable or Mythos for harder jobs, the split Codema covered in the Opus 5.5 builders guide. How that API spend stacks up against other frontier models is the kind of comparison in Codema’s LLM API pricing roundup. Our AI models explained guide has the current ChatGPT, Claude and Gemini list prices in one place.

Anthropic’s prerelease claims include formalizing the proof of Fermat’s Last Theorem in 30 hours with a team of about 1,000 Fable 5.2 agents, and solving two open problems in black-hole physics. Those are Anthropic’s company claims from its launch post, not independently verified results. The same post also says Fable 5.2 designed three working electronic prototypes with Diode Computers and improved robotics evaluations, which are also Anthropic's own reported results. Builders who already wire Claude into agent workflows can read Codema’s Claude Code subagents vs skills vs hooks guide for how the tooling layer around a model often matters as much as the model ID.