Anthropic Unveils Claude Fable 5.2 and Mythos 5.2 with Faster Coding, Reasoning and Vision

Anthropic’s new Claude Fable 5.2 and Mythos 5.2 models promise clearer communication, faster responses and better performance on technical tasks. They’re now cheaper and available across major platforms, with extra usage for paid subscribers.

Anthropic Unveils Claude Fable 5.2 and Mythos 5.2 with Faster Coding, Reasoning and Vision

Anthropic announced the launch of two upgraded AI models—Claude Fable 5.2 and Claude Mythos 5.2—on November 9, 2026. Both models deliver clearer output, up to 30% faster response times, and improved abilities in coding, reasoning and vision compared with their 5.1 predecessors. The upgrades also bring lower token pricing and new safeguards, making them attractive for developers, researchers and enterprise users.

Key takeaways

  • Fable 5.2 and Mythos 5.2 generate outputs up to 30% faster than 5.1 versions.
  • Token pricing drops to $8 per million input tokens (output $40 per million).
  • Fable 5.2 uses fewer tokens on challenging tasks when set to medium effort.
  • Both models excel in coding, multidisciplinary reasoning, vision and cybersecurity tasks.
  • Extra Fable usage this week for eligible paid Anthropic plans.

What’s new in Fable 5.2 and Mythos 5.2

Anthropic says the two models share the same underlying architecture but differ in safety guardrails. Fable 5.2 includes cybersecurity and biology safeguards, while Mythos 5.2 is offered through the company’s Life Sciences and Cyber Verification Programs and Claude Security on Enterprise plans.

  • Speed: Output generation is up to 2.5 × faster in fast‑mode, with overall response times roughly 30% quicker than 5.1.
  • Token efficiency: On medium‑effort settings, Fable 5.2 consumes about 55% fewer “blocks” (a proxy for token usage) than its 5.1 predecessor.
  • Performance gains: Benchmarks show both models outperform the previous Opus 5.5 in computer use, knowledge work, cybersecurity, agentic scientific research, agentic coding and multidisciplinary reasoning.

Pricing and availability

Both models are available today on all Anthropic platforms, including Claude Code, the Claude Platform, Google Cloud, and Microsoft Foundry. Pricing has been reduced for Fable 5.2 to $8 per million input tokens (output $40 per million), making it cheaper than competing GPT‑6 Astra.

ModelInput token priceOutput token price
Claude Fable 5.2$8 / M tokens$40 / M tokens
Claude Mythos 5.2Price not disclosed (enterprise‑only)—

How to choose the right model

Anthropic recommends starting with Opus 5.5 for most coding and knowledge tasks. When a project requires more ambitious, open‑ended technical work, users can step up to Fable 5.2 or Mythos 5.2. For cost‑conscious workloads, the smaller Sonnet 5.5 or Haiku 5.5 models remain viable options.

Typical use cases

  • Complex coding projects: Fable 5.2 can write, debug and even migrate codebases, as Anthropic reports a 26‑fold speed boost when moving a service from Python to Rust.
  • Scientific research: Both models excel in agentic scientific tasks, with Mythos 5.2 offering specialized life‑science safeguards.
  • Vision tasks: Improved image understanding helps with document analysis, design review and multimodal applications.

What this means for you

If you already have a paid Anthropic subscription, you may receive one to three extra Fable resets this week, and Pro‑plan users get up to 50% of their weekly usage limits counted toward Fable usage. To take advantage, simply log in to your Anthropic account and check the usage dashboard for any added credits.

Developers can enable the fast‑mode on Claude Code or the Claude Platform to benefit from the 2.5 × token‑generation speed boost. Enterprise customers needing zero‑data‑retention (ZDR) can activate the Enterprise Frontier Safeguards system, which now supports Fable 5.2.

What happens next

Anthropic plans to continue refining the Fable and Mythos families, with future releases expected to further reduce latency and expand safety features. The company also indicated ongoing work to improve evaluation methods, as the new models raise challenges for real‑world testing.

Users should monitor Anthropic’s release notes for upcoming updates, especially if you rely on the models for high‑stakes or regulated workloads. Keeping an eye on pricing changes and the availability of extra usage credits can help you manage costs while experimenting with the newer capabilities.

Key takeaways

  • Claude Fable 5.2 and Mythos 5.2 are up to 30% faster and more token‑efficient than 5.1.
  • Fable 5.2 price drops to $8 / M input tokens, making it cheaper than GPT‑6 Astra.
  • Both models shine in coding, reasoning, vision and cybersecurity tasks.
  • Extra Fable usage credits are available this week for eligible paid plans.
  • Enterprise users can enable zero‑data‑retention safeguards on Fable 5.2.

Frequently asked questions

How much does Claude Fable 5.2 cost per million tokens?

$8 for input tokens and $40 for output tokens.

Is Claude Mythos 5.2 available for individual developers?

Mythos 5.2 is offered through Anthropic’s Life Sciences, Cyber Verification Programs and Enterprise plans.

What performance improvements do the new models have?

They generate output up to 30% faster and use fewer tokens on challenging tasks.

Can I get extra usage for Fable 5.2 on my current plan?

Yes, eligible paid plans receive one to three extra Fable resets and up to 50% of weekly limits counted toward Fable usage.

Do the new models support zero data retention?

Fable 5.2 can be used with Anthropic’s Enterprise Frontier Safeguards for zero‑data‑retention policies.

Sources

This story was prepared with the help of AI tools, based on the reporting listed above. Details of developing stories may change.