Claude Mythos 5.2
Claude Fable 5.2 with fewer safeguards, for organizations in Anthropic's verification programs
Context window
1M
Input / 1M tokens
$8.00
Output / 1M tokens
$40.00
Provider
Anthropic
Same prices as Claude Fable 5.2 and 20% below Mythos 5.1: cache reads $0.20 per 1M, 5-minute cache writes $10, 1-hour cache writes $16, Batch API $4/$20. Fast mode costs $16/$80 per 1M and is limited to organizations in Anthropic's verification programs. · Data verified 2026-11-11
Claude Mythos 5.2 (November 9, 2026) is the same model as Claude Fable 5.2, with the same specifications and $8/$40 pricing, but with fewer of the cybersecurity and biology safeguards that block some Fable requests. It replaced Claude Mythos 5.1 and is offered only to organizations verified through Anthropic's Cyber Verification Program or Life Sciences Verification Program. Claude Enterprise customers can also use it through Claude Security, Anthropic's codebase vulnerability scanner, without joining the Cyber Verification Program. In Anthropic's CVE Rediscovery evaluation (1,829 known CVEs), Mythos 5.2 found 89.6% versus 78.3% for Mythos 5.1 and 81.3% for Opus 5.5 without safeguards.
How to access it
Restricted to organizations verified through Anthropic's Cyber Verification Program (cybersecurity) or Life Sciences Verification Program (biomedical research); apply to the program that covers your use case or contact your account team. Approved organizations call it on the Claude API (claude-mythos-5-2), Amazon Bedrock (anthropic.claude-mythos-5-2), Claude Models on Google Cloud and Microsoft Foundry; Claude Platform on AWS does not list it. Claude Enterprise customers can use it inside Claude Security without joining the Cyber Verification Program.
Strengths
- ✓Same model as Fable 5.2 without the cyber and biology classifiers that block some Fable requests
- ✓CVE Rediscovery 89.6% vs 78.3% for Mythos 5.1 (Anthropic-reported)
- ✓Terminal-Bench 4.0 70.5% without safeguards (Anthropic system card)
- ✓De novo protein binder design score of 85.1 vs 79.1 for Mythos 5.1
- ✓$8/$40 per 1M, 20% below Mythos 5.1
Best for developers who...
When to choose it (and when not to)
Reach for Claude Mythos 5.2 when...
- →Your organization is approved through Anthropic's Cyber or Life Sciences Verification Program
- →Defensive security or biomedical work that Fable 5.2's safeguards block
- →You use Mythos 5.1 today: 5.2 costs 20% less and scored higher on Anthropic's cyber evaluations
Look elsewhere if...
- ✕Your organization is not verified through one of Anthropic's access programs
- ✕General-purpose work: Claude Fable 5.2 is the same model with standard safeguards
- ✕You need it on Claude Platform on AWS, which does not list it
- ✕You need zero data retention without Anthropic's authorization (30-day retention applies)
How to use it
- ›The API default effort is high; use xhigh for long vulnerability-research runs
- ›Use the 1M context window to review a full codebase in one request
Quickstart
Pythonimport anthropic
client = anthropic.Anthropic()
message = client.messages.create(
model="claude-mythos-5-2",
max_tokens=16000,
messages=[{"role": "user", "content": "Your prompt here"}],
)
print(next(block.text for block in message.content if block.type == "text"))Works only for organizations approved through Anthropic's Cyber or Life Sciences Verification Program. On Amazon Bedrock the ID is anthropic.claude-mythos-5-2.
API model id: claude-mythos-5-2
Benchmarks
| Benchmark | Score | Notes |
|---|---|---|
| Terminal-Bench 4.0 | 70.5% (max effort) | Anthropic system card, evaluated without safeguards, 660 trials; Fable 5.2 scored 67.6% with safeguards |
| CVE Rediscovery | 89.6% | Anthropic-reported, October 2026 run on 1,829 CVEs; Mythos 5.1 78.3%, Opus 5.5 81.3% without safeguards |
| CyScenarioBench | 76.3% | Anthropic-reported; Mythos 5.1 61.7%, Opus 5.5 67.6% without safeguards |
| Terminal-Bench-Science 0.1 | 75.3% | Anthropic system card, without safeguards; Fable 5.2 74.6% |
| De novo binder design (24 hr) | 85.1 | Anthropic-reported protein design score; Mythos 5.1 79.1, Opus 5.5 82.6 without safeguards |
Compare Claude Mythos 5.2
Compare Claude Mythos 5.2 with any other model
Build a comparison →All model comparisons →