What Is Claude Mythos? Mythos vs Fable vs Opus Explained

Claude Mythos is Anthropic’s most capable class of AI models, a tier that sits above Opus. With the release of Claude Fable 5.2 and Claude Mythos 5.2, many people are now meeting three names at once and wondering which one they can actually use. The short version: Fable and Mythos are the same model, Fable is the version with extra safety guards that paying Claude users can try, Mythos is kept for vetted security and science teams, and Opus remains the everyday workhorse.

Quick answer: Mythos is Anthropic’s top model class. Fable is the same model with strong cybersecurity and biology safeguards, available on paid Claude plans and the API. Mythos without those safeguards goes only to vetted organizations through Anthropic’s verification programs and Project Glasswing. Opus is the tier below, and Anthropic recommends it as the starting point for most coding and knowledge work. The Free plan gets neither Fable nor Opus.

Mythos vs Fable vs Opus at a glance

What is Claude Mythos?

Anthropic describes Mythos on its Claude Mythos page (external site) as its most capable model class for cybersecurity and biology research. In the model lineup, Mythos-class models sit above Opus, which in turn sits above Sonnet and Haiku.

The name first appeared with Claude Mythos Preview, an unreleased research model that Anthropic shared only with security partners. Since then Anthropic has released numbered Mythos versions (Mythos 5, 5.1 and now 5.2), each paired with a Fable model of the same version number.

What makes Mythos different is not a secret extra skill but the safeguards. Very capable models can help defenders find and fix software flaws, and they can help researchers in biology. The same abilities could be misused, so Anthropic keeps the version with fewer limits in those two areas for organizations it has checked.

Mythos vs Fable: same model, different safeguards

This is the part that confuses most people. In Anthropic’s own words, Claude Fable and Claude Mythos are the same underlying model. The difference is what each one is allowed to do:

• Fable includes strong cybersecurity and biology safeguards. When a request falls into those sensitive areas, Fable hands it to an Opus model instead of answering itself. Because of these guards, Anthropic can offer Fable to everyone on paid plans.
• Mythos is the same model without certain cybersecurity or biology safeguards. It is only for vetted organizations doing beneficial work in those areas, such as security teams defending software or researchers in the life sciences.

So for everyday work, writing, coding, analysis or research outside those sensitive topics, Fable and Mythos behave the same. Anthropic even reports benchmark scores for Fable and says any gap with Mythos comes from tasks where Fable’s safeguards step in.

Two practical differences for organizations: Mythos keeps conversations for 30 days by default for safety monitoring (Anthropic says enterprise safeguards can make it compatible with zero data retention), and Anthropic’s Claude Security product for enterprises runs on Mythos.

Mythos vs Opus: which tier do you need?

Opus is the tier below Mythos and Fable, and for most people it is still the right choice. Anthropic itself recommends starting with Claude Opus 5.5, which it calls its strongest daily driver, for most coding and knowledge tasks, and moving up to Fable or Mythos only when the job is big enough to need it.

Anthropic says Opus 5.5 performs at the level of the earlier Fable 5.1 on most work, and it is the first Opus model to launch with a similar class of safeguards to Fable. Where Fable pulls ahead is on long, complex, open-ended projects: Anthropic says the longer and more complex the task, the bigger Fable’s lead over its other models.

A simple way to choose:

• Pick Opus for everyday writing, coding, research, spreadsheets and agent tasks. It is included in paid plan limits and is the cheaper of the two on the API.
• Pick Fable for the hardest technical work: large codebase changes, long multi-step projects, tough science or finance problems where Opus falls short. Anthropic also suggests Fable as an “advisor” or “orchestrator” that hands simpler pieces to smaller models.
• Mythos is not a choice most people can make. Your organization has to apply and be approved.

If you are new to Claude, our guide on how to use Claude AI covers the basics, and Sonnet on the Free plan is a good place to start before paying for anything.

Who can use Mythos, Fable and Opus?

Here is how access works on each Claude plan, based on Anthropic’s plans page (external site) and its help center article on Fable models on your plan (external site):

On Max, Ultra and premium seats, once you have used Fable for half of your weekly limit you can keep going with usage credits or switch to another model. For the monthly price of each plan and how the plans compare with ChatGPT and Gemini, see our ChatGPT vs Claude vs Gemini plans comparison.

How to get access to Mythos

Individuals cannot sign up for Mythos. Anthropic offers it to vetted organizations through two routes:

1. Cyber Verification Program: security teams and cybersecurity professionals can apply for advanced cyber capabilities on Mythos and other Claude models.
2. Life Sciences Verification Program: life sciences researchers can apply for advanced capabilities in biology.

Both are listed on Anthropic’s Mythos page. If you just want the most capable model you can use today, Fable on a paid plan is the answer.

What is Project Glasswing?

Project Glasswing (external site) is Anthropic’s effort to help defenders secure the world’s most important software before attackers can use AI to break it. It started with Claude Mythos Preview, which Anthropic says found thousands of high-severity vulnerabilities, including in every major operating system and web browser.

Instead of releasing that model to the public, Anthropic gave access to a group of launch partners, Amazon Web Services, Apple, Broadcom, Cisco, CrowdStrike, Google, JPMorganChase, the Linux Foundation, Microsoft, NVIDIA and Palo Alto Networks, plus more than 40 other organizations that build or maintain critical software. The program was later extended to more organizations. Anthropic also committed up to $100 million in usage credits for the research preview and $4 million in donations to open-source security groups.

Anthropic says it does not plan to make Mythos Preview generally available. Glasswing members have since moved into the Specialized Access tier of the Cyber Verification Program. For readers, the takeaway is simple: Glasswing is why your operating system, browser and other apps may get security fixes for bugs that AI helped find, so keep automatic updates on.

Is Claude Mythos safe, and should you worry?

Anthropic built the Fable and Mythos split precisely because these models are strong at finding software flaws. The safeguards in Fable (and now in Opus 5.5) are meant to block the most dangerous uses, while Mythos without those limits goes only to checked organizations. For everyday users, nothing changes in how you stay safe: update your devices, use strong passwords and two-factor authentication, and be careful with links and attachments.

Curious how AI that takes actions on its own works? Our plain-English guide to agentic AI explains the basics, and our roundup of the best AI tools for coding shows where Claude fits among developer tools.

FAQ

Is Claude Mythos available to the public?

No. Mythos is only for vetted organizations through Anthropic’s verification programs and Project Glasswing. The public version of the same model is Claude Fable, available on paid Claude plans and the API.

Are Claude Mythos and Fable the same?

Yes, they are the same underlying model. Fable adds strong cybersecurity and biology safeguards; Mythos lifts some of them for approved security and life sciences work.

Is Claude Fable free?

No. Fable is not on the Free plan. Pro users pay for it with usage credits, while Max, Ultra and premium Team and Enterprise seats can use it for up to half of their weekly limit.

Is Mythos better than Opus?

Mythos-class models are Anthropic’s most capable, and the lead grows on long, complex tasks. For most everyday work, though, Anthropic recommends starting with Opus, which is included in paid plans and costs less on the API.

How do I apply for Claude Mythos?

Organizations can apply through Anthropic’s Cyber Verification Program (for security teams) or Life Sciences Verification Program (for researchers). Individuals cannot sign up on their own.