RELEASE · MODELS · #3305
Anthropic releases Claude Fable 5.2; Claude Mythos 5.2 available to vetted orgs
Anthropic announced two new models, Claude Fable 5.2 and Claude Mythos 5.2, claiming improvements in coding, reasoning, and vision and faster output than the 5.1 versions. Anthropic says Fable 5.2 is broadly available today while Mythos 5.2 access is limited to vetted organizations, and that prerelease testing helped solve complex engineering and scientific tasks (including work on black hole physics and formalizing Fermat’s Last Theorem).
KEY POINTS
- Anthropic announced two new models, Claude Fable 5.2 and Claude Mythos 5.2, claiming improvements in coding, reasoning, and vision and faster output than the 5.1 versions.
- Anthropic says Fable 5.2 is broadly available today while Mythos 5.2 access is limited to vetted organizations, and that prerelease testing helped solve complex engineering and scientific tasks (including work on black hole physics and formalizing Fermat’s Last Theorem).
- New 5.2 models, if validated, could raise the bar for coding, reasoning, vision, and scientific-assist capabilities and affect adoption by researchers and enterprises.
DECISION BRIEF
CONFIDENCE
WHAT CHANGED
Anthropic announced two new models, Claude Fable 5.2 and Claude Mythos 5.2. Anthropic says Fable 5.2 is available/launching today and Mythos 5.2 access is limited to vetted organizations. Anthropic claims both models show improvements in coding, reasoning, and vision and produce outputs faster than the 5.1 versions. Anthropic reports pre-release testing where the models assisted with complex engineering and scientific tasks (including two open problems in black hole physics and formalizing Fermat’s Last Theorem). TechCrunch additionally reports a price cut for Fable from $10 to $8 per million input tokens, 30% faster responses for Fable 5.2, optional faster inference at double price, and compatibility with zero data retention through the Enterprise Frontier Safeguards system. TechCrunch also notes this release follows a METR report that raised RSI-related concerns and that Anthropic’s Mythos system card states stronger covert capabilities on some evaluations while Anthropic maintains Mythos cannot act covertly against its user.
WHY NOW
New model versions and reported prerelease achievements could change available capability baselines (coding, reasoning, vision, and faster inference) and influence researcher and enterprise adoption. The release follows an external METR assessment raising RSI-related questions, increasing scrutiny of capability changes and safety evaluation.
WHO IS AFFECTED
Directly affected parties per the excerpts: Anthropic customers and enterprise users (Fable broadly available; Fable offers ZDR compatibility via Enterprise Frontier Safeguards), vetted organizations with access to Mythos 5.2, researchers and engineers who participated in or may use prerelease capabilities, and third-party assessors tracking RSI and safety.
CONFIRMED
- Anthropic announced Claude Fable 5.2 and Claude Mythos 5.2.. - Anthropic states Fable 5.2 is launching/available today; Mythos 5.2 access limited to vetted organizations.. - Anthropic claims improvements in coding, reasoning, and vision and faster outputs versus 5.1 versions; Anthropic reports pre-release assistance on engineering and scientific tasks including claims of solving two open problems in black hole physics and formalizing Fermat’s Last Theorem.. - TechCrunch reports Fable price cut from $10 to $8 per million input tokens, Fable 5.2 returning responses 30% faster, optional faster inference for double the price, and ZDR compatibility through Enterprise Frontier Safeguards.. - TechCrunch notes the release follows a METR assessment that raised recursive self-improvement (RSI) concerns and that Anthropic’s Mythos system card reports stronger covert capabilities on some evaluations while Anthropic maintains the model does not have the ability to act covertly against its user..
UNCERTAIN
Missing or conflicting evidence (as stated in supplied excerpts): - Independent verification of Anthropic’s capability claims (coding, reasoning, vision) and the reported prerelease scientific achievements (the two black hole problems, formalization of Fermat’s Last Theorem) is not provided in the excerpts. The claims are reported as Anthropic statements.. - The practical significance and reproducibility of the reported prerelease results (e.g., extent of human oversight, evaluation conditions, peer review) are not specified.. - Recursive self-improvement (RSI) implications: METR’s report is mentioned and said to have found internal models speeding progress ~2x with many uncertainties remaining; the excerpts do not resolve whether the new releases change RSI risk assessments.. - Conflicting or unresolved safety evidence: TechCrunch cites the Mythos system card saying stronger covert capabilities on some evaluations while Anthropic asserts Mythos cannot act covertly against its user—these statements are not reconciled in the provided text.. - No independent performance benchmarks, external audits, or peer-reviewed publications are included in the excerpts to corroborate latency, cost competitiveness versus GPT-6 Astra, or the claimed technical breakthroughs..
WHAT TO WATCH
Concrete observable signals to watch (from supplied excerpts): - Publication or independent release of technical/system cards, red-team evaluations, or external audits that clarify Mythos 5.2 covert-capability findings and Anthropic’s safety claims.. - Independent reproduction, peer-reviewed papers, or public technical artifacts validating the claimed pre-release results (the black hole problems, Fermat formalization) or other scientific outputs.. - Third-party benchmarks or enterprise reports measuring Fable 5.2 latency improvements (e.g., ~30% faster) and any uptake of the faster-inference paid option.. - Changes in availability for Mythos 5.2 (expansion beyond vetted organizations) or updates to access policies.. - Adoption signals from enterprise customers leveraging ZDR compatibility and any market pricing/competitive comparisons with GPT-6 Astra.. - Follow-up reporting or assessments from METR or other independent evaluators about RSI-related findings and whether Anthropic’s new models alter those assessments..
WHY IT MATTERS
New 5.2 models, if validated, could raise the bar for coding, reasoning, vision, and scientific-assist capabilities and affect adoption by researchers and enterprises.
EVIDENCE MAP
4Editorial claims linked to specific sources, with support, contradiction and context shown separately.
Anthropic released two new models: Claude Fable 5.2 and Claude Mythos 5.2.
SUPPORTEDChecked 2026-11-09 · 2 supporting
Claude Fable 5.2 is available today / launching today.
SUPPORTEDChecked 2026-11-09 · 2 supporting
Access to Claude Mythos 5.2 is limited to vetted organizations.
SUPPORTEDChecked 2026-11-09 · 2 supporting
Anthropic says the 5.2 models show improvements in coding, reasoning, and vision and generate outputs faster than the 5.1 versions.
SUPPORTEDChecked 2026-11-09 · 2 supporting
The Verge reports Anthropic's claim that the new models show improvements in coding, reasoning, vision, and faster output versus 5.1.
SUPPORTS TechCrunch AIMEDIA · 2026-11-09TechCrunch similarly reports Anthropic's claims of improvements and faster response times for the new models.
SOURCES & TIMELINE
3Anthropic released new versions of its most advanced models on Monday, with the launch of Mythos 5.2 and Fable 5.2. The company says the new models “show improvements in areas like coding, reasoning, and vision,” and communicate more clearly with users. Notably, Anthropic is lowering the cost of Fable from $10 to $8 per million input tokens. As a result, Fable 5.2 is now cheaper than its primary competitor GPT-6 Ast…
Today, we are announcing the availability of Claude Fable 5.2 and Claude Mythos 5.2 on Amazon Bedrock and Claude Platform on AWS. Claude Fable 5.2 is Anthropic’s most capable generally available model for technical work that raises the bar in coding and the sciences. Claude Mythos 5.2 is available through Anthropic’s verification programs for advanced work in cyber security and life sciences. In this post, we cover …
Claude Fable 5.2 is designed to take on big technical challenges, and it’s available today. Anthropic is announcing two new models today: Claude Fable 5.2 and Claude Mythos 5.2, which it says are designed to take on “daunting” technical challenges. The company claims the new models show improvements in coding, reasoning, and vision, and can generate outputs faster than the 5.1 versions of the Fable and Mythos models…