Not established from the available sources.
ANTHROPIC · MODEL RELEASE TRACKER
Claude Mythos 5.1
Claude Mythos 5.1 is a named Anthropic model. The supplied article states Anthropic has an internal, unreleased research model that it says is more capable and better aligned than Mythos 5.1.CURRENT SNAPSHOT1/5 DIMENSIONS WITH DATA
The dimensions that change the decision.
Not established from the available sources.
Not established from the available sources.
Not established from the available sources.
Access restored to AISIAccess for the UK’s AI Security Institute (AISI) to Claude Mythos 5.1 was restored after the White House reversed its block.
Previously accessible via Project Glasswing (trusted access)Anthropic states that Project Glasswing had given a group of organizations securing critical software trusted access to Claude Mythos.
RELEASE TIMELINE
Published, source-backed release events only.
VERIFIABLE FACTS
Every value stays attached to a source and date.
AVAILABILITY · Access restored to AISIDEVELOPER CLAIM
Access for the UK’s AI Security Institute (AISI) to Claude Mythos 5.1 was restored after the White House reversed its block.
AVAILABILITY · Previously accessible via Project Glasswing (trusted access)DEVELOPER CLAIM
Anthropic states that Project Glasswing had given a group of organizations securing critical software trusted access to Claude Mythos.
SAFETY · Alignment compared to Anthropic internal modelDEVELOPER CLAIM
Anthropic says its internal, unreleased model is better aligned than Mythos 5.1.
CAPABILITIES · Listed among Anthropic's 'most capable models'DEVELOPER CLAIM
Anthropic lists Claude Mythos 5.1 among its most capable models that are made available to certain security programs.
RELEASE · Release statusDEVELOPER CLAIM
Described in reporting as a new model.
LIMITATIONS · Access restrictedDEVELOPER CLAIM
Anthropic restricted access to Claude Mythos 5.1.
CAPABILITIES · Outperformed by Anthropic internal modelDEVELOPER CLAIM
Anthropic says an internal, unreleased research model “shows capabilities greater than Mythos 5.1.”
WHAT CHANGED
Stored passport versions, without reconstructed history.
SAFETY · Alignment compared to Anthropic internal model+ Anthropic says its internal, unreleased model is better aligned than Mythos 5.1.
CAPABILITIES · Outperformed by Anthropic internal model+ Anthropic says an internal, unreleased research model “shows capabilities greater than Mythos 5.1.”
SAFETY · Access associated with advanced cyber capabilities and reduced blocking classifiers for qualifying security teams− The expanded CVP is described as making advanced cyber capabilities and reduced blocking classifiers available to qualifying security professionals; Claude Mythos 5.1 is included among the models available in those tiers.
LIMITATIONS · Access restricted+ Anthropic restricted access to Claude Mythos 5.1.
AVAILABILITY · Access restored to AISI+ Access for the UK’s AI Security Institute (AISI) to Claude Mythos 5.1 was restored after the White House reversed its block.
AVAILABILITY · Available via expanded Cyber Verification Program (CVP) access tiers− Claude Mythos 5.1 is listed among Anthropic’s 'most capable models' and is made available to qualifying security professionals through the expanded Cyber Verification Program (CVP).
RELEASE · Release status+ Described in reporting as a new model.
AVAILABILITY · Availability− Reportedly available only to U.S. organizations.
AVAILABILITY · Available via expanded Cyber Verification Program (CVP) access tiers+ Claude Mythos 5.1 is listed among Anthropic’s 'most capable models' and is made available to qualifying security professionals through the expanded Cyber Verification Program (CVP).
CAPABILITIES · Listed among Anthropic's 'most capable models'+ Anthropic lists Claude Mythos 5.1 among its most capable models that are made available to certain security programs.
AVAILABILITY · Previously accessible via Project Glasswing (trusted access)+ Anthropic states that Project Glasswing had given a group of organizations securing critical software trusted access to Claude Mythos.
SAFETY · Access associated with advanced cyber capabilities and reduced blocking classifiers for qualifying security teams+ The expanded CVP is described as making advanced cyber capabilities and reduced blocking classifiers available to qualifying security professionals; Claude Mythos 5.1 is included among the models available in those tiers.