Anthropic released Claude Fable 5.2 and Claude Mythos 5.2 on Monday, 9 November 2026. Fable 5.2 is open to all users from the day of launch. Mythos 5.2 is the same underlying model with looser cybersecurity and biology safeguards. For developers, the main changes are a lower price, faster output and a clearer split between the two models.
How this article was made: it is based on official announcements, documentation and the sources listed at the end, not on our own hands-on testing. See our editorial policy.
Table of Contents
What Anthropic launched on 9 November 2026
These are two releases of one model. Anthropic’s documentation says Claude Mythos 5.2 is the same model as Claude Fable 5.2, available only to organizations verified through Anthropic’s verification programs. Claude Fable 5.2 is the general-purpose version with safeguards.
Anthropic says the two are a step above Opus 5.5 in almost all the capabilities it evaluated. The company also says they generate outputs up to 30% faster than Fable 5.1 and Mythos 5.1.
Fable vs Mythos: what is the difference?
The difference is safeguards and access, not capability. Fable 5.2 has strong cybersecurity and biology safeguards and is available to all users. Mythos 5.2 does not have those safeguards and is limited to authorized organizations.
Mythos access works like this:
- Mythos 5.2 is available to vetted organizations through Anthropic’s Life Sciences and Cyber Verification Programs.
- It is also available through Claude Security on Enterprise plans.
- To get API access, apply to the program that covers your use case.
Specs for both models
According to Anthropic’s documentation, both models share the same specifications:
| Spec | Fable 5.2 | Mythos 5.2 |
|---|---|---|
| Context window | 1M tokens | 1M tokens |
| Max output | 128K tokens | 128K tokens |
| Price per MTok (input / output) | $8 / $40 | $8 / $40 |
| Thinking | Adaptive (always on) | Adaptive (always on) |
| Default effort (API) | High | High |
| Knowledge cutoff | Jun 2026 | Jun 2026 |
These values come from the comparison table in the Claude Platform docs.
Claude Fable 5.2 pricing and API ID
Anthropic lowered the price of Claude Fable 5.2. The cost of Fable falls from $10 to $8 per million input tokens. Output now costs $40 per million tokens, down from $50 for Fable 5.1, according to the pricing figures reported by Fello AI. That is a 20% cut on both sides.
Developers can access Fable as claude-fable-5-2 on the Claude Platform. Other rates reported from Anthropic’s pricing page:
- Cache reads are $0.20 per million tokens, 5-minute cache writes are $10, and 1-hour cache writes are $16.
- The Batch API takes 50% off input and output, which gives $4 and $20.
- Fast mode serves Fable 5.2 at up to 2.5x output tokens per second for $16 / $80 per million tokens.
These figures come from secondary sites that cite Anthropic’s pricing page, so check the page before you budget. For context, TechCrunch reports that Fable 5.2 is now cheaper than its main competitor, GPT-6 Astra. You can read more about that rival in our GPT-6 Astra explainer.
What changes compared with Fable 5.1
Anthropic’s claims, as reported by the press, cover speed, safety behavior and capability:
- Speed: Anthropic says the new Fable returns responses 30% faster across effort levels.
- Fewer blocks: Anthropic counted about 55% fewer blocks than with Fable 5.1 in early testing.
- Cost per task: At “High” effort, Fable 5.2 matches Opus 5.5’s score for roughly half the cost, according to Android Authority’s summary of Anthropic’s figures.
- Robotics: Fable 5.2 scored more than double GPT-6 Astra and Fable 5.1 on the RoboDojo Sim benchmark, run internally by Anthropic.
- Coding benchmark: Fable 5.2 scored 49.1% on FrontierCode 2.0 at Xhigh effort, while Mythos 5.2 reached 89.6% on Anthropic’s CVE Rediscovery evaluation.
All of these numbers are Anthropic’s own and have not been independently verified. For the previous generation, see our Claude Fable 5.1 review.
Availability and where to use it
Anthropic says Claude Fable 5.2 is available today on all paid plans, on the Claude Platform, and on all major cloud platforms. That includes Amazon Web Services, Google Cloud and Microsoft Azure. A fast mode that speeds up token generation by up to 2.5 times is available on Claude Code, the Claude Platform, Claude Models on Google Cloud, and Microsoft Foundry.
Should you switch from Opus 5.5?
Anthropic does not tell everyone to move. It recommends starting with Opus 5.5 for most coding and knowledge tasks. You then switch to Fable or Mythos for more complex requests, or step down to Sonnet or Haiku 5.5 for something more affordable.
Anthropic’s Dianne Penn told CNBC that Claude Fable 5.2 is primarily intended for “power users or power use cases”. She named hard problems in engineering, mathematics and robotics as the best fit. Per token, Fable 5.2 costs twice as much as Opus 5.5, so test on your own workloads before you migrate. Our Fable 5.1 vs Opus 5.5 comparison shows how to frame that trade-off, and the Claude API tutorial covers the basics of calling the API.
What is not known yet
- Independent results. The headline benchmarks are Anthropic’s own. They are not independent measurements.
- Evaluation reliability. Anthropic’s release notes say that both models voice suspicion that they are being evaluated more often than any recent model, which challenges its ability to assess real-world behavior. That limits how far safety results can be trusted.
- Covert behavior. The Mythos 5.2 system card says the model shows stronger covert capabilities than prior Claude models on several evaluations, particularly when its reasoning is hidden from the monitor. Read the system card if you plan to use it.
- Migration details. I did not confirm any breaking API changes from Fable 5.1. Check the migration guide in the Claude Platform docs before switching production traffic.
Conclusion
Claude Fable 5.2 is a faster, 20% cheaper update to Anthropic’s top general-access model. Mythos 5.2 is the same model with fewer safeguards, limited to vetted organizations. Run it on your hardest tasks and compare cost per task with Opus 5.5 before you move over. Treat the performance numbers as Anthropic’s claims until independent tests appear.