AI model · Anthropic
Mythos 5.2, Explained: Access, Pricing, Safeguards
Mythos 5.2 is the same model as Claude Fable 5.2 with lighter cyber and biology safeguards, released November 9, 2026 for vetted organizations only, at $8 per million input tokens. Synexa does not serve it; use Synexa when your app also needs images, video, audio or 3D from one API key.
- Verification programs only
- $8 in / $40 out per MTok
- Synexa: image, video, audio, 3D APIs
What is Mythos 5.2?
Mythos 5.2 is Anthropic's most capable model for cybersecurity and life-sciences research, released on November 9, 2026. It is not a separate, bigger model. Anthropic states that Claude Fable 5.2 and Mythos 5.2 are the same underlying model: Fable ships with strong cyber and biology safeguards so anyone can use it, while Mythos relaxes those safeguards for organizations Anthropic has verified.
That split matters in practice. Fable's classifiers block penetration testing, exploit generation, binary vulnerability work and most research biology, and a blocked query is answered by an older Opus model instead. Mythos is the version a vetted security team or lab uses when that work is the job.
Outside security and biology, the two behave the same: Anthropic publishes one benchmark table for both, with separate Mythos rows only where Fable's safeguards would intervene. For the general-purpose side, read our Fable 5.2 guide.
Mythos 5.2 specs: IDs, context and limits
Published figures as of November 2026. Anthropic lists the same specifications and price as Fable 5.2.
- Maker
- Anthropic
- Released
- November 9, 2026
- Status
- Active, verification required
- Claude API ID
claude-mythos-5-2- Amazon Bedrock ID
global.anthropic.claude-mythos-5-2(bedrock-runtime)- Other platforms
- Claude Models on Google Cloud (Google or Anthropic managed), Microsoft Foundry
- Input / output
- Text and images in, text out
- Context window
- 1M tokens
- Max output
- 128K tokens
- Thinking
- Adaptive, always on; default effort
high - Knowledge cutoff
- June 2026
- Retirement
- Not sooner than November 9, 2027
- Data retention
- 30 days by default for safety monitoring
Mythos 5.2 vs Fable 5.2 vs Mythos 5.1
Same model as Fable, different doors. Mythos 5.1, the September 2026 version, was limited to a small group of vetted partners.
Who can use it
- Mythos 5.2
- Verified organizations (and individuals at Defense Access)
- Fable 5.2
- Everyone on paid plans and the API
- Mythos 5.1
- A small set of vetted organizations
Cyber and biology safeguards
- Mythos 5.2
- Relaxed for approved work
- Fable 5.2
- Strong; block most offensive security and research biology
- Mythos 5.1
- Relaxed for approved work
Price per MTok
- Mythos 5.2
- $8 / $40
- Fable 5.2
- $8 / $40
- Mythos 5.1
- $10 / $50
Fast mode
- Mythos 5.2
- $16 / $80, verification programs only
- Fable 5.2
- $16 / $80
- Mythos 5.1
- not compared
Context / max output
- Mythos 5.2
- 1M / 128K
- Fable 5.2
- 1M / 128K
- Mythos 5.1
- 1M / 128K
Claude Platform on AWS
- Mythos 5.2
- Not listed
- Fable 5.2
- Yes
- Mythos 5.1
- not compared
| Mythos 5.2 | Fable 5.2 | Mythos 5.1 | |
|---|---|---|---|
| Who can use it | Verified organizations (and individuals at Defense Access) | Everyone on paid plans and the API | A small set of vetted organizations |
| Cyber and biology safeguards | Relaxed for approved work | Strong; block most offensive security and research biology | Relaxed for approved work |
| Price per MTok | $8 / $40 | $8 / $40 | $10 / $50 |
| Fast mode | $16 / $80, verification programs only | $16 / $80 | not compared |
| Context / max output | 1M / 128K | 1M / 128K | 1M / 128K |
| Claude Platform on AWS | Not listed | Yes | not compared |
"Not listed" means the platform does not appear on Anthropic's Mythos 5.2 model page. "Not compared" means we did not check that item for Mythos 5.1.
Mythos 5.2 pricing and an example bill
Mythos 5.2 pricing matches Fable 5.2: $8 per million input tokens and $40 per million output tokens on the Claude API, 20% below Mythos 5.1. Prices as of November 2026.
Input, per MTok
- Mythos 5.2
- $8
- Mythos 5.1
- $10
Output, per MTok
- Mythos 5.2
- $40
- Mythos 5.1
- $50
5-minute cache write
- Mythos 5.2
- $10
- Mythos 5.1
- not compared
1-hour cache write
- Mythos 5.2
- $16
- Mythos 5.1
- not compared
Cache read
- Mythos 5.2
- $0.20
- Mythos 5.1
- $0.25
Batch API
- Mythos 5.2
- 50% off input and output
- Mythos 5.1
- not compared
Fast mode
- Mythos 5.2
- $16 / $80
- Mythos 5.1
- not compared
Example: 100K in + 10K out
- Mythos 5.2
- $1.20
- Mythos 5.1
- $1.50
| Rate | Mythos 5.2 | Mythos 5.1 |
|---|---|---|
| Input, per MTok | $8 | $10 |
| Output, per MTok | $40 | $50 |
| 5-minute cache write | $10 | not compared |
| 1-hour cache write | $16 | not compared |
| Cache read | $0.20 | $0.25 |
| Batch API | 50% off input and output | not compared |
| Fast mode | $16 / $80 | not compared |
| Example: 100K in + 10K out | $1.20 | $1.50 |
The example row is our arithmetic from the list prices: 0.1M × $8 plus 0.01M × $40 = $1.20. The same call through the Batch API is $0.60.
How to get Mythos 5.2 access
There is no sign-up button. Mythos 5.2 access goes through one of Anthropic's verification programs, or through Claude Security on Enterprise.
Cyber Verification Program
Life Sciences Verification Program
Claude Security on Enterprise
How does Claude Mythos 5.2 score?
Anthropic shows separate Mythos results only where Fable's safeguards would get in the way. Opus 5.5's cyber and protein-design scores are without safeguards; its Terminal-Bench-Science score is with safeguards on. We have not run these ourselves.
CyScenarioBench (cybersecurity)
- Mythos 5.2
- 76.3%
- Mythos 5.1
- 61.7%
- Opus 5.5
- 67.6%
CVE Rediscovery (cybersecurity)
- Mythos 5.2
- 89.6%
- Mythos 5.1
- 78.3%
- Opus 5.5
- 81.3%
De novo binder design, 24 h (protein design)
- Mythos 5.2
- 85.1
- Mythos 5.1
- 79.1
- Opus 5.5
- 82.6
Terminal-Bench-Science 0.1
- Mythos 5.2
- 75.3%
- Mythos 5.1
- 52.6% (Fable 5.1)
- Opus 5.5
- 58.7%
| Benchmark | Mythos 5.2 | Mythos 5.1 | Opus 5.5 |
|---|---|---|---|
| CyScenarioBench (cybersecurity) | 76.3% | 61.7% | 67.6% |
| CVE Rediscovery (cybersecurity) | 89.6% | 78.3% | 81.3% |
| De novo binder design, 24 h (protein design) | 85.1 | 79.1 | 82.6 |
| Terminal-Bench-Science 0.1 | 75.3% | 52.6% (Fable 5.1) | 58.7% |
Vendor-reported scores at each model's highest available effort. Anthropic also notes that on its own vulnerability-identification evals the model is roughly on par with Opus 5.5, so the gap depends on the task.
What changed in Mythos 5.2
The headline gains are in finding and reproducing vulnerabilities, plus price and behavior.
More bugs found
Better calibration
Lower price
Stronger alignment scores
Fewer sandbox escapes
Science results
Mythos 5.2 access rules to know first
Approval is per organization and per use, and a few API rules changed from Mythos 5.1.
- Apply from the exact organization and account where you will run the work
- Keep each approved tier in its own workspace, role or cloud account
- Plan for 30-day retention unless you already hold a ZDR exemption
- Use the
computer_toolset_20260801toolset for computer use
- Serve customers with it before Anthropic approves that external use
- Edit the system prompt or tools and then replay later thinking blocks; that returns a 400
- Expect Pro plans to include it; Pro pays for Mythos with usage credits
- Assume Bedrock access is simple; it currently needs a data retention exemption
Can you run Mythos 5.2 on Synexa?
No. Synexa does not serve Mythos, Fable or any other Claude model, and it offers no chat LLMs; its only text outputs are LLaVA-13B image answers, Whisper speech transcripts and an NSFW image-check label. For the language side, go through Anthropic or one of the clouds above.
Synexa covers the media half of an AI product through one REST API. One key gives you image models such as Nano Banana Pro at $0.10 a request, music, speech, upscaling, and image-to-3D models such as Hunyuan3D 3.1 at $0.16 a request. Every price is listed on its model page in Explore, and new accounts start with $0.10 of free credit. The quickstart has copy-paste calls for Python and Node, and each model page shows a plain HTTP example.
- Pay per request, no subscription
- Python and Node clients, plus plain HTTP
- Image, video, audio, upscaling and 3D in one catalog
Synexa models to pair with a Claude agent
If a Claude agent plans the work, these Synexa models can produce the assets. Flat prices per request, as listed on each model page in November 2026.
Product or report imagery
- Synexa model
- Nano Banana Pro
- Price per request
- $0.10
Text-heavy graphics
- Synexa model
- GPT Image 2
- Price per request
- $0.10
Textured 3D model from one image
- Synexa model
- Hunyuan3D 3.1
- Price per request
- $0.16
Fast draft 3D mesh
- Synexa model
- TRELLIS
- Price per request
- $0.02
Transcribe a recorded briefing
- Synexa model
- Whisper v3
- Price per request
- $0.004
Upscale a screenshot or photo
- Synexa model
- Real-ESRGAN
- Price per request
- $0.0015
| Task | Synexa model | Price per request |
|---|---|---|
| Product or report imagery | Nano Banana Pro | $0.10 |
| Text-heavy graphics | GPT Image 2 | $0.10 |
| Textured 3D model from one image | Hunyuan3D 3.1 | $0.16 |
| Fast draft 3D mesh | TRELLIS | $0.02 |
| Transcribe a recorded briefing | Whisper v3 | $0.004 |
| Upscale a screenshot or photo | Real-ESRGAN | $0.0015 |
These are Synexa's own catalog prices, not Anthropic's. Synexa does not run any Claude model.
Who should use Mythos 5.2, and who should not?
Mythos 5.2 fits a narrow group: security teams validating exploits or running authorized penetration tests, vulnerability researchers, and life-science labs whose work trips Fable's biology filters. If Fable 5.2 already does your work without blocks, Mythos adds nothing, since it is the same model at the same price.
For most coding and knowledge work, Anthropic still recommends starting on Claude Opus 5.5 and moving up only when needed; see our Claude Opus 5.2 explainer for why there is no Opus 5.2. Teams that also generate images, voice or 3D can compare media APIs in our WaveSpeed AI alternative page, or read about another recent model, Gemini 4 Hydrogen.
Related pages
More model explainers and API comparisons.
Frequently asked questions
What is Mythos 5.2?
Mythos 5.2 is Anthropic's model for cybersecurity and life-sciences research, released November 9, 2026. It is the same underlying model as Claude Fable 5.2, with the cyber and biology safeguards relaxed for verified organizations. It reads text and images, writes text, holds a 1M-token context and returns up to 128K tokens per response.
What is the difference between Claude Mythos 5.2 and Fable 5.2?
Only the safeguards and who can use it. Both are one model with the same specs and the same $8 / $40 price. Fable 5.2 is open to everyone but blocks penetration testing, exploit generation, binary vulnerability work and most research biology. Mythos 5.2 lifts those limits for organizations approved through Anthropic's Cyber or Life Sciences verification programs.
How much does Mythos 5.2 cost?
On the Claude API it costs $8 per million input tokens and $40 per million output tokens as of November 2026, down from $10 and $50 on Mythos 5.1. Cache reads are $0.20 per million tokens, the Batch API halves the token price, and fast mode costs $16 and $80 for verified organizations. A call with 100K tokens in and 10K out costs about $1.20.
How do I get access to Mythos 5.2?
Apply to Anthropic's Cyber Verification Program in its Verification Portal, or to the Life Sciences Verification Program for biology work. The cyber program needs a paid Claude plan or a billed Console account, the biology one an Enterprise, Team or API account. Individuals can apply only for cyber Defense Access. Claude Enterprise customers can also use it inside Claude Security without joining the program.
Can individuals use Claude Mythos?
Yes, at one tier. Independent researchers, open-source maintainers and bug bounty hunters can apply for Defense Access on a Pro, Max, Ultra, Team, Enterprise or API plan, and that access covers only themselves. Red Team Access and Specialized Access, for offensive testing and critical infrastructure, are for organizations only and require Team, Enterprise or API plans.
Is Mythos 5.2 included in Claude Max?
Yes, once your account is approved. Mythos models follow the same usage rules as Fable: on Max and Ultra, Mythos and Fable share up to 50% of the weekly usage limit, and that usage also counts toward your regular limits. On Pro, Mythos is available only with usage credits. The Claude Free plan cannot use it at all.
Does Mythos 5.2 keep my data?
By default, yes, for 30 days, so Anthropic can monitor for misuse across many requests. Eligible Cyber Verification Program customers who already have zero data retention can keep it for this model. Anthropic's Enterprise Frontier Safeguards, in private beta, runs that monitoring on data stored in the customer's own infrastructure instead.
Can I use Mythos 5.2 through Synexa?
No. Synexa does not serve any Claude model or any chat language model. Synexa is a pay-per-request API for generative media: images, video, speech and music, upscaling and image-to-3D. Teams often let a Claude agent plan and write, then call Synexa to render the image, clip, voice-over or 3D model. New Synexa accounts get $0.10 of free credit.
Pair Mythos 5.2 with a media API
Keep Mythos for the security or research reasoning, and generate the images, audio and 3D models your product needs on Synexa. Pay per request, start with $0.10 free credit.