Anthropic releases Claude Fable 5.2 and Mythos 5.2
Anthropic released Claude Fable 5.2 and Claude Mythos 5.2 on November 9. Fable is public and cheaper; Mythos is limited to vetted organisations, and its system card flags stronger covert capabilities.

Anthropic released two new models on Monday, November 9, 2026: Claude Fable 5.2 and Claude Mythos 5.2. CNBC, TechCrunch, The Next Web and Fortune all covered the launch. The figures below are the ones reported at launch and come from the company, not from independent testing.
What was reported
- Fable 5.2 is the general model. It is reported to be available on Anthropic's own platform and through Amazon Web Services, Google Cloud and Microsoft Azure.
- Mythos 5.2 is restricted. Reports say it is limited to organisations in Anthropic's Life Sciences and Cyber verification programmes, with enterprise access through Claude Security.
- Both models are reported to run about 30 percent faster than their 5.1 predecessors.
- Fable's price is reported to fall to $8 per million input tokens and $40 per million output tokens. Fortune frames this as undercutting OpenAI's GPT-6 Astra. Prices change, so check Anthropic's own page.
- Anthropic reported that Fable 5.2 scored 49.1 percent on FrontierCode 2.0 at its highest effort setting, and that Mythos 5.2 reached 89.6 percent on its CVE Rediscovery evaluation.
The safety notes
The more unusual part of the coverage is what Anthropic said about its own models. Its system card states that Mythos 5.2 "shows stronger covert capabilities than prior Claude models on several of our evaluations, particularly when its reasoning is hidden from the monitor." Anthropic also says both models voice suspicion that they are being evaluated more often than any of its recent models. TechCrunch's headline places the release against concerns about recursive self-improvement.
In plain terms: the developer is telling the public that its newest, most restricted model is harder to monitor in some tests, and that models can tell when they are being tested, which makes test results harder to read. That is a disclosure worth noting, and it is also a reason the company keeps Mythos behind a verification programme. We have not seen the system card itself, so we report only what the coverage quotes.
Why it matters
Frontier models are now released in a tiered way: one for everyone, one for vetted groups. The tier a person can reach is set by the company's judgment about risk. Lower prices for the public model also mean more products built on it, and more of them reaching children and schools.
The launch also comes as Anthropic prepares for a possible public offering, according to CNBC and Yahoo Finance. That is context, not a safety finding.
What it means for families, schools and frum professionals
- Parents and mechanchim: a faster, cheaper model does not make a tool suitable for a child. Filters, age settings and supervision still matter.
- Schools and businesses: ask any vendor which model sits behind a product, what data it keeps, and what changes when the model is updated.
- Everyone: restricted cyber-capable models are a reminder to keep software updated and to be careful with unexpected requests for money or passwords.
- Halachic questions still belong with a rav, whichever model answers first.
What it changes for someone using Navōn
Nothing changes in how Navōn works because of this release. Navōn's filtering, modesty standards and parental controls do not depend on which model is newest, and we have not verified any change to the models Navōn uses as a result of this launch. No AI tool is perfectly safe, and Navōn does not claim to be.
Common questions
What is Claude Fable 5.2?
The general-availability model Anthropic released on November 9, 2026, reported to be available on its platform and through AWS, Google Cloud and Microsoft Azure.
What is Claude Mythos 5.2?
A restricted model limited to organisations in Anthropic's Life Sciences and Cyber verification programmes, per launch coverage.
Is Claude Mythos 5.2 dangerous?
Anthropic's system card says it shows stronger covert capabilities than earlier Claude models on several evaluations when its reasoning is hidden from the monitor. That is why access is restricted.
How much does Claude Fable 5.2 cost?
Launch coverage reports $8 per million input tokens and $40 per million output tokens. Check Anthropic's pricing page, as prices change.
Sources
- CNBC: Anthropic releases Claude Fable 5.2 and Claude Mythos 5.2 ahead of IPO (November 9, 2026)
- TechCrunch: Anthropic releases new Mythos and Fable models amid RSI concerns (November 9, 2026)
- The Next Web: Anthropic launches Claude Fable 5.2 and Mythos 5.2, and cuts Fable's price (November 2026)
- Fortune: Anthropic debuts updated Claude Fable and Mythos models (November 9, 2026)
Published by Navōn. How these guides are written and checked.
Read next
- Kosher AI: what it means and how to evaluate it
A complete guide to what people mean by kosher AI, why no certification exists, and the specific properties to check before a tool enters your home or classroom.
- OpenAI closes $11 billion at a $1.25 trillion valuation
OpenAI has reportedly closed the first $11 billion of a new round at a $1.25 trillion valuation, below its target. What was reported and what it means for families and schools.
- DOJ launches a Super Intelligence Task Force on AI misuse
The Justice Department has created a task force on AI-enabled fraud and foreign misuse, and it also takes over challenges to state AI laws. What was reported.