🔍 Intelligence Synthesis JP/EN
Ethics · Regulation · Technology — Pharma Practice Notes

AI Tech Flash ── 2026-11-10

Anthropic Releases Fable 5.2 and Mythos 5.2

Two models, one base; Mythos is vetted-access only

What changes

Anthropic is offering two products built on the same base model, split by safety tier. Fable 5.2 is for general use. Mythos 5.2 is for vetted organizations.

According to The News International, Fable 5.2 is meant for tasks that run autonomously for hours or days. Examples include migrations across repositories and formal mathematical proofs. A repository is a unit in which program code is managed.

Why it matters

Shattered.io and Tech Insider cite CNBC as reporting that Anthropic called these its first frontier models since CEO Dario Amodei urged slowing AI development.

Anthropic says both models went through rigorous safety and alignment testing. It also says it built new evaluations and post-launch monitoring tools before release. Shattered.io notes that these claims have not been independently verified.

What it is

Claude Fable 5.2 and Claude Mythos 5.2 are two AI models that Anthropic released on November 9, 2026. According to The News International, they share the same underlying architecture and differ in safety tier and deployment scope.

According to Tech Insider, Anthropic calls them its "most capable models for daunting technical work." Shattered.io, however, says it could not confirm benchmarks.

Who and where

Anthropic built them. According to Tech Insider, Fable 5.2 is available on paid Claude plans, in Claude Code and on Claude Platform. It is also on AWS, Google Cloud and Microsoft Azure. Its model ID is claude-fable-5-2.

Mythos 5.2 is not sold generally. Tech Insider reports it is reserved for vetted cybersecurity and life-sciences users in a program called Project Glasswing.

Why it was built

According to The News International, there is a growing need for AI that can work through multi-step, ill-defined, long-running tasks and catch its own errors. Fable 5.2 was built for that kind of work.

Mythos 5.2 targets high-risk research in cybersecurity and life sciences, including penetration testing and protein research, per the same report. Because it could be misused, access is limited.

How it works

According to The News International, both models have a 1-million-token context window. A token is a unit used to count text, and the context window is the most text a model can read at once. The report says this lets them take in entire codebases in one run.

According to Tech Insider, Anthropic says both generate tokens up to 30% faster than their Claude 5.1 predecessors. Prices are $8 per million input tokens and $40 per million output tokens. The Batch API, which returns results asynchronously, costs half that: $4 and $20.

Key points

  1. Anthropic released Fable 5.2 and Mythos 5.2 on November 9, 2026.
  2. Fable 5.2 is for general use; Mythos 5.2 is for vetted organizations.
  3. Pricing is $8 input and $40 output per million tokens; speed is up to 30% faster, Anthropic says.

Conclusion

Anthropic released two models on one base, separated by safety tier. Both are priced the same.

According to shattered.io, benchmarks and test results are unconfirmed. Independent verification is still needed.

Quotes and sources

“Fable 5.2 is designed as Anthropic's flagship, deeply deliberative model for rigorous day-to-day engineering, science, and mathematics.”
── The News International, 2026-11-10
“Anthropic says both models generate tokens up to 30% faster than their Claude 5.1 predecessors.”
── Tech Insider, 2026-11-10
“Mythos 5.2 does not go on general sale.”
── Tech Insider, 2026-11-10
“Anthropic released Claude Fable 5.2 and Claude Mythos 5.2 on November 9, 2026”
── shattered.io, 2026-11-10

Source articles

Written by AI (Claude) from the cited articles, then checked by Jev (TypeSafe's judgement model), a Gemini fact-check and a verbatim check of every quote.