Claude Fable and Mythos 5.2: What Launched, Who Gets Which, and the New Price

Anthropic released Claude Fable and Mythos 5.2 on November 9, 2026, and cut the price by 20% to $8 per million input tokens and $40 per million output tokens. Fable 5.2 is open to paying customers. Mythos 5.2 is the same model with fewer safeguards, and only verified organizations can use it.

This is the third release in the Fable line since June. We covered the first in Claude Fable 5 and the second in Claude Fable 5.1. The split between the public model and the gated one has not changed. The price, the speed and the number of blocked requests have.

The short version: most agencies and nonprofit IT teams will use Fable 5.2, not Mythos 5.2. It costs less per token than Fable 5.1, but it still requires 30-day data retention. On the API it also rejects some requests that worked on 5.1. Pro and Team standard users get a free week of Fable through November 16. Test now, and do not switch production code until you have checked the breaking changes.

This post is the overview. Tomorrow we publish separate deep dives on Fable 5.2 and on Mythos 5.2.

What Anthropic launched on November 9

According to Anthropic’s launch announcement, Fable 5.2 and Mythos 5.2 are the same underlying model. Fable 5.2 carries strong cybersecurity and biology safeguards, so it is available to all users. Mythos 5.2 drops some of those safeguards and is limited to vetted organizations.

Anthropic says both models write output up to 30% faster than Fable 5.1 and Mythos 5.1. It also says testers found both communicate more clearly than earlier Fable and Mythos models.

Anthropic’s Fable product page calls Fable 5.2 its most capable model for technical work, with coding, finance and the sciences named first. On Anthropic’s own FrontierSWE v2 coding test, Fable 5.2 scores 69.9%, against 62.3% for Claude Opus 5.5. These are vendor numbers. We will look at them in tomorrow’s Fable 5.2 post.

Nothing was retired. Anthropic’s model deprecations page still lists Fable 5, Fable 5.1, Mythos 5 and Mythos 5.1 as active, with no retirement before June 2027 for the 5.0 models and September 2027 for 5.1.

Fable and Mythos 5.2 at a glance

Here is how Fable and Mythos 5.2 compare, based on Anthropic’s API release notes and its What’s new in Claude Fable 5.2 page.

Fable 5.2 still routes some requests to older models. Anthropic’s announcement says blocked cyber requests get an answer from Claude Opus 4.8, and blocked biology requests get one from Claude Opus 5. Anthropic says Fable 5.2 blocks about 55% fewer requests overall than Fable 5.1.

Who gets Fable and Mythos 5.2

The Fable product page lists Fable 5.2 on the Pro, Max, Ultra, Team and Enterprise plans. How much you can use depends on the plan.

On Max, Ultra, Team premium seats and premium seats on legacy seat-based Enterprise plans, Fable is part of the plan. On Pro and Team standard seats, Fable normally runs on usage credits. We explained that split in our post on Fable 5 usage limits.

Mythos 5.2 is a different matter. The announcement names three routes:

• The Cyber Verification Program, for security professionals and organizations doing defensive security work.
• The Life Sciences Verification Program, for researchers doing beneficial biomedical work.
• Claude Security, which includes Mythos 5.2 for Claude Enterprise customers, even outside the cyber program.

Anthropic’s Cyber Verification Program article says organizations apply once through the Verification Portal and Anthropic aims to respond within seven business days. As of launch night, that article still lists Mythos 5.1 in its access tiers, not 5.2. It also says approval does not switch access on; an admin has to assign it.

On cloud platforms, the What’s new page lists Mythos 5.2 on Amazon Bedrock, both Google Cloud offerings and Microsoft Foundry. It does not list Mythos 5.2 on Claude Platform on AWS.

Fable and Mythos 5.2 prices

Anthropic’s pricing page gives both models the same rates. The release notes confirm Fable 5.1 and Mythos 5.1 cost $10 and $50, so the cut is 20% on input and output.

The pricing page says the full 1M-token window bills at the standard rate. US-only inference costs 1.1 times the standard rate. Fast mode doubles the price for up to 2.5 times faster output, per the release notes.

The cache line matters most if you run long agent loops. Our earlier post on Fable 5.1 cost and cache hit rates shows why, and our prompt caching explainer covers the basics.

Price comparisons with other vendors are circulating in press coverage. The Anthropic pages we checked do not compare its prices with OpenAI or Google, so check those rates on each vendor’s own pricing page.

How to try Claude Fable 5.2

In the Claude apps, pick Fable 5.2 from the model picker. Anthropic’s launch offers article says your default model does not change. Anthropic’s announcement says medium effort is the default in the apps and in Claude Code.

The launch offers break down by plan:

• Pro and Team standard seats: every Fable model is included from November 9 through November 16 at 11:59 pm PT, up to 50% of weekly usage limits, plus one Fable limit reset that expires the same night.
• Max, Ultra, Team premium and legacy Enterprise premium seats: one to three Fable limit resets, based on how long you have paid. They expire March 8, 2027.
• Not included: the Free plan, standard seats on legacy seat-based Enterprise, self-serve Enterprise, usage-based Enterprise and HIPAA-ready Enterprise.

Resets go only to accounts with paid billing in place before 10:00 am PT on November 9. You use one from Settings, then Usage. A reset restores usage. It does not raise your limits.

On the API, call claude-fable-5-2. The API default effort is high, per the What’s new page.

What agencies and nonprofit IT should do about Fable and Mythos 5.2

Start with the model you need. Anthropic itself recommends starting with Claude Opus 5.5 for most coding and knowledge work, and moving up to Fable for the hardest open-ended jobs. We covered that model in Claude Opus 5.5.

Then work through this list:

1. Pro and Team standard users: try Fable 5.2 on real work before November 16. The free week ends then.
2. Max, Ultra and Team premium users: check Settings, then Usage, for your resets. Note the March 8 expiry.
3. API teams: rerun your evaluations on claude-fable-5-2 and compare cost per task, not just cost per token.
4. Compliance owners: confirm that 30-day retention is acceptable for the data you send. Zero data retention needs Anthropic’s express approval.
5. Security clients: if a client does defensive security work, they can apply for the Cyber Verification Program now.

For context on how often the Fable rules have changed, see our posts on the June suspension and the July restore.

Similar moves at OpenAI and Google

Anthropic’s Fable page compares Fable 5.2 with OpenAI’s GPT-6 Astra and Google’s Gemini 4 Carbon on its benchmarks. We covered OpenAI’s model lineup in which OpenAI model to use and Google’s model in Gemini 4 Argon. Check each vendor’s own pages for current prices.

Frequently Asked Questions