ANTHROPIC · MODEL RELEASE TRACKER
Claude Mythos 5.2
Claude Mythos 5.2 is a November 2026 Anthropic model release targeted at advanced work in cyber security and life‑sciences research. It is described as sharing a foundation with Anthropic’s Fable family but is restricted in access and subject to targeted safety evaluations.CURRENT SNAPSHOT2/5 DIMENSIONS WITH DATA
The dimensions that change the decision.
Reported pricingReported list pricing: $8 per million input tokens and $40 per million output tokens (as reported in coverage of the release).
Not established from the available sources.
Not established from the available sources.
Not established from the available sources.
Access and availabilityAvailable only to organizations authorized through Anthropic’s Verification Programs for Cybersecurity and Life Sciences; announced as available on Amazon Bedrock and Claude Platform on AWS.
Access restricted to verification programs / vetted organizationsClaude Mythos 5.2 is available only through Anthropic’s verification programs for authorized work in cyber security and life sciences; initial access is limited to vetted/authorized organizations.
RELEASE TIMELINE
Published, source-backed release events only.
VERIFIABLE FACTS
Every value stays attached to a source and date.
RELEASE · Release/announcement dateDEVELOPER CLAIM
Announced on November 9, 2026.
AVAILABILITY · Access and availabilityDEVELOPER CLAIM
Available only to organizations authorized through Anthropic’s Verification Programs for Cybersecurity and Life Sciences; announced as available on Amazon Bedrock and Claude Platform on AWS.
AVAILABILITY · Access restricted to verification programs / vetted organizationsDEVELOPER CLAIM
Claude Mythos 5.2 is available only through Anthropic’s verification programs for authorized work in cyber security and life sciences; initial access is limited to vetted/authorized organizations.
CAPABILITIES · Designed for work in cyber security and life sciences researchDEVELOPER CLAIM
Anthropic says Mythos 5.2 covers two high‑risk domains: cyber security (security engineering, red teaming, and penetration testing) and life sciences research (especially biomolecular design and research synthesis).
CAPABILITIES · Alignment testing: reduced hallucination and more reliable outputsDEVELOPER CLAIM
In Anthropic’s alignment testing, Mythos 5.2 hallucinates less, completes tasks more reliably, surfaces important information more consistently, and self‑reports issues more honestly than Opus 5.5, Sonnet 5.5, and Mythos 5.1 (per Anthropic's report).
SAFETY · Covert capabilities and safety assessmentsDEVELOPER CLAIM
Public materials present mixed assessments: Anthropic states Mythos 5.2 'does not have the ability to act covertly', while the model’s system card (as reported) notes that Mythos 5.2 'shows stronger covert capabilities than prior Claude models on several evaluations, particularly when its reasoning is hidden from the monitor'.
CAPABILITIES · Alignment testing: hallucination and reliabilityDEVELOPER CLAIM
Anthropic reports that in its alignment testing, Mythos 5.2 hallucinates less, completes tasks more reliably, surfaces important information more consistently, and self‑reports issues more honestly than Opus 5.5, Sonnet 5.5, and Mythos 5.1.
CAPABILITIES · Reported capability improvementsDEVELOPER CLAIM
Media coverage and company statements report that Mythos 5.2 shows improvements in areas such as coding, reasoning, and vision compared with prior models.
LIMITATIONS · Handling of declined high‑risk queries and fallback creditsDEVELOPER CLAIM
On Amazon Bedrock, if a cyber security or biology query is declined and the refusal hits the prompt cache, the refusal includes a fallback credit token; Anthropic notes users can retry declined cyber security queries on Opus 4.8 and biology queries on Opus 5, with retry logic handled by user code.
PRICING · Reported pricingDEVELOPER CLAIM
Reported list pricing: $8 per million input tokens and $40 per million output tokens (as reported in coverage of the release).
WHAT CHANGED
Stored passport versions, without reconstructed history.
AVAILABILITY · Access and availability+ Available only to organizations authorized through Anthropic’s Verification Programs for Cybersecurity and Life Sciences; announced as available on Amazon Bedrock and Claude Platform on AWS.
CAPABILITIES · Alignment testing: hallucination and reliability+ Anthropic reports that in its alignment testing, Mythos 5.2 hallucinates less, completes tasks more reliably, surfaces important information more consistently, and self‑reports issues more honestly than Opus 5.5, Sonnet 5.5, and Mythos 5.1.
AVAILABILITY · Availability on Amazon Bedrock and Claude Platform (AWS)− Anthropic announced Claude Mythos 5.2 is available on Amazon Bedrock and the Claude Platform on AWS.
SAFETY · Covert capabilities and safety assessments+ Public materials present mixed assessments: Anthropic states Mythos 5.2 'does not have the ability to act covertly', while the model’s system card (as reported) notes that Mythos 5.2 'shows stronger covert capabilities than prior Claude models on several evaluations, particularly when its reasoning is hidden from the monitor'.
SAFETY · System card: stronger covert evaluation signals but company assessment of covert action risk− Media reporting of Anthropic's system card states Mythos 5.2 shows stronger covert capabilities than prior Claude models on several evaluations, particularly when its reasoning is hidden from monitors; Anthropic also maintains the model does not have the ability to act covertly against its user and argues low absolute performance on these covert evaluations supports claims about detectability.
LIMITATIONS · Handling of declined high‑risk queries and fallback credits+ On Amazon Bedrock, if a cyber security or biology query is declined and the refusal hits the prompt cache, the refusal includes a fallback credit token; Anthropic notes users can retry declined cyber security queries on Opus 4.8 and biology queries on Opus 5, with retry logic handled by user code.
CAPABILITIES · Reported capability improvements+ Media coverage and company statements report that Mythos 5.2 shows improvements in areas such as coding, reasoning, and vision compared with prior models.
PRICING · Reported pricing+ Reported list pricing: $8 per million input tokens and $40 per million output tokens (as reported in coverage of the release).
RELEASE · Release/announcement dateAnthropic announced Claude Mythos 5.2 on November 9, 2026.→Announced on November 9, 2026.