Key takeaways
- Claude Mythos 5.2 is open only to vetted organizations, and security teams apply through the Cyber Verification Program.
- Anthropic says CVP decisions arrive by email within about seven business days, and approval must then be assigned to a workspace, role or linked cloud account.
- Life sciences teams register interest in a separate beta program, and Claude Enterprise customers can reach Mythos 5.2 through Claude Security.
You can't sign up for Claude Mythos 5.2 the way you sign up for a chat plan. Anthropic gives it only to vetted organizations, and the main way in is the Cyber Verification Program, which you apply to through Anthropic's Verification Portal. This guide covers claude mythos 5.2 access for security teams, life sciences teams and Claude Enterprise customers, as of November 9, 2026.
Anthropic announced Mythos 5.2 on November 9, together with Claude Fable 5.2. The two are the same underlying model. Fable has stronger cybersecurity and biology safeguards, so anyone can use it. Mythos is the version with those safeguards relaxed, for organizations that do dual-use work in those areas. Our news article on the launch covers the model itself. This page covers how to get in.
Who can get Claude Mythos 5.2
Anthropic names three routes:
| If you are | Route | What you get |
|---|---|---|
| A cybersecurity professional | Apply to the Cyber Verification Program (CVP) | Access to Mythos 5.2, in one of three tiers |
| A life sciences researcher | Register interest in the Life Sciences Verification Program | Mythos, Opus and Sonnet models with safeguards that are more permissive for biology work |
| A Claude Enterprise customer | Use Claude Security | Claude Security runs on the full model family, including Mythos 5.2, even if you are not in CVP |
If you only want the strongest model that everyone can use, that is Fable 5.2, and it needs no application. Anthropic still recommends Claude Opus 5.5 for most everyday coding and knowledge work.
How to apply to the Cyber Verification Program
CVP gives qualifying security professionals reduced blocking on advanced cyber tasks. Access comes in three tiers: Defense, Red Team and Specialized. Anthropic's help center says generally available models stay open to everyone for tasks like secure code review, threat modeling and patching. Malware analysis and exploit validation may still be blocked without CVP.
Who applies depends on who you are:
- Organizations apply once. Individuals inside an organization don't apply separately. Admins assign seats.
- Independent researchers, maintainers and bug bounty hunters apply as individuals, but only for Defense Access, and they must be on a paid plan.
The steps, from the help center:
- Sign in to the Verification Portal, go to Programs, choose Cyber Verification Program and select Apply.
- Give organization and applicant details so Anthropic can verify your identity.
- Describe your security work.
- Attest to the security controls required for the tier you want.
- Wait for the decision. Anthropic places you at the highest tier your information supports and emails the decision within about seven business days.
Anthropic looks at your organization's defense focus, whether your identity can be verified, your legal and regulatory environment, and the risk of the access being diverted. Organizations that mainly serve military, intelligence or law enforcement customers get a more cautious review. Higher tiers have stricter requirements, and Specialized Access includes an in-depth review with the US government.
An application can be turned down. The help center lists these reasons: your main business falls outside defense, verification fails, your operating environment is a concern, your customers lean toward military, intelligence or law enforcement, or you asked for a higher tier than you qualify for.
What to do after you're approved
Approval alone doesn't switch anything on. You have to assign the access to a workspace, a role or a linked cloud account. How you do that depends on where you use Claude:
| Where you use Claude | What to do |
|---|---|
| Claude Console or API | Create a workspace (the default workspace can't be used) and add the approved users |
| Claude Enterprise | Assign the program to a custom role |
| Team, Pro, Max, Ultra | Nothing. It attaches at the organization level |
| Claude Platform on AWS, Google Cloud, Microsoft Foundry | Link your cloud account in the portal first, then turn on access for it. Foundry needs a Claude model deployed beforehand |
| Amazon Bedrock | Only for customers with Enterprise Frontier Safeguards |
| Other third-party platforms | Get an enrollment link from the platform. Defense and Red Team access only |
On cloud platforms, Anthropic says Mythos access typically trails approval by about five business days.
Anthropic's announcement and help center don't give a developer model ID for Mythos 5.2. The model ID in the announcement is claude-fable-5-2, for Fable. Check the model list in your workspace once access is on.
Plan limits and price
Mythos 5.2 costs the same as Fable 5.2: $8 per million input tokens and $40 per million output tokens. Anthropic also sells a faster mode at $16 and $80, and says fast mode for Mythos is limited to Verification Program organizations.
On consumer and team plans, the help center says:
- Pro: Mythos models need usage credits, and you must turn them on.
- Max and Ultra: Mythos and Fable share one allowance of up to 50% of your weekly usage limit. After that you can continue with usage credits or switch models.
Life Sciences Verification Program
This program is in beta. It is open to verified organizations on Claude Enterprise and Team plans, and on the Claude Platform (API). Free, Pro, Max and Ultra plans are not eligible right now.
An organization admin fills in a form to register interest. Anthropic then checks eligibility. Submitting the form doesn't guarantee access. If your organization qualifies, Anthropic follows up with details on how to apply formally. The page doesn't say what the form asks, how long the review takes or what the verification criteria are, so direct questions go to your account team.
Rules that come with access
- The Usage Policy still applies in full. Running the models on your own code is allowed.
- Anthropic can review, narrow or withdraw a grant.
- CVP requires data retention so Anthropic can monitor for misuse. Organizations with a data retention exemption for Fable or Mythos models can use CVP with zero data retention for covered models.
- Defense Access has until December 15, 2026 to move to phishing-resistant multi-factor authentication and stop using API keys. Until then, some MFA is required and API keys expire every seven days.
- If you want to use CVP to deliver work to your own customers, you need separate approval for each service or product. Managed services, where your staff run the model on customer-owned code and customers only receive deliverables, are open for applications. Products that customers use directly aren't accepting applications yet.
- Anthropic says people already in Project Glasswing or CVP don't need to reapply. They move to the new tiers under their current terms.
What we don't know yet
Anthropic hasn't published a wait time beyond the seven business days for CVP decisions, or any acceptance numbers. It also hasn't said how long Mythos 5.2 will stay gated. Check the help center article before you apply, because the tiers and security requirements can change.
Frequently asked questions
Can anyone use Claude Mythos 5.2?
No. Anthropic says Mythos 5.2 is available to vetted organizations that work in dual-use areas such as cybersecurity and life sciences. Claude Fable 5.2, the same model with stronger safeguards, is open to all users.
How long does Cyber Verification Program approval take?
Anthropic's help center says decisions are emailed within about seven business days. On cloud platforms, Mythos access typically trails approval by about five business days.
Is Claude Mythos 5.2 free?
No. Anthropic lists $8 per million input tokens and $40 per million output tokens, the same as Fable 5.2. On Pro, Mythos models need usage credits turned on.
Can an individual get Claude Mythos 5.2 access?
Independent researchers, maintainers and bug bounty hunters can apply as individuals, but only for Defense Access, and they must be on a paid plan. Staff at an organization don't apply separately, because admins assign seats.
Sources
- Claude Fable 5.2 and Mythos 5.2Anthropic
- Cyber Verification ProgramAnthropic Help Center
- Life Sciences Verification ProgramAnthropic




