OpenAI and Anthropic Restrict AI Model Access Amid U.S. Cybersecurity Review
Summary
OpenAI is limiting GPT-5.6 Sol to administration-approved customers, while Anthropic received approval for limited Mythos 5 access by cyber defenders and infrastructure providers amid U.S. review of advanced AI models’ cybersecurity risks.
Key points
- OpenAI said GPT-5.6 Sol will initially be available only to customers approved by the Trump administration, with broader availability anticipated in the coming weeks.
- Anthropic said the government approved limited redeployment of Mythos 5 to cyber defenders and infrastructure providers after restrictions were lifted.
- Anthropic had taken Mythos 5 and Fable 5 offline to comply with a directive blocking their use by foreign nationals.
- OpenAI said Sol is better at helping users find and fix vulnerabilities than carry out cyberattacks, while acknowledging potential unforeseen risks when combined with other tools.
- The administration’s AI oversight framework allows federal review of advanced systems’ national security risks for up to 30 days before public release, but the framework is not yet fully developed.
- Anthropic said it will continue working with the government to expand Mythos access and make Fable available again to general users.
Article Details
- Event Type
- Government cybersecurity review and access restrictions for new AI models: the Trump administration restricted or approved limited releases of models from OpenAI and Anthropic.
- Impact
- GPT-5.6 Sol was limited to customers approved by the administration, while Mythos 5 was approved for redeployment to a small group of cyber defenders and infrastructure providers. Anthropic said Fable 5 remained unavailable; the article reports no actual cyberattack or compromise resulting from the models.
People
Alex StamosCybersecurity expert who said he had reviewed Anthropic's Fable research analysis and disputed that the government's action was supported by evidence of risks distinct from those of other available models.Dario AmodeiAnthropic CEO involved in discussions with the administration; Anthropic warned that Mythos could find software flaws that might be weaponized, and Amodei was involved in the company's response to government restrictions.David SacksCo-leads the president's council of technology and science advisers and said Anthropic CEO Dario Amodei had described Mythos as a cyber weapon, while acknowledging the model had advanced cyber capabilities.Donald TrumpAs president, signed an executive order establishing a federal review framework for advanced AI systems and issued a directive that led Anthropic to take models offline.Howard LutnickU.S. Commerce Secretary involved in negotiations with AI industry executives and who wrote to Anthropic that its work addressing government concerns had yielded significant progress.Lori TrahanU.S. representative and co-author of a bipartisan AI regulation bill who criticized the administration's company-by-company decisions on access to new models.Sam AltmanOpenAI CEO who spoke with U.S. Commerce Secretary Howard Lutnick about the model release.
Vendors
AnthropicIts chief rival, Anthropic, announced hours later that the Trump administration has approved a limited release of its strongest cybersecurity model, two weeks after the U.S.OpenAIChatGPT maker OpenAI said Friday it is restricting the release of its new artificial intelligence model at the request of President Donald Trump’s administration, the latest in an unprecedented government vetting of AI
Products
ClaudeOpenAI’s staggered release of a powerful new AI system follows actions the government took earlier this month against Anthropic, maker of the Claude chatbot.Fable 5Anthropic took offline two new AI models, known as Fable 5 and Mythos 5, just days after unveiling them to comply with a Trump directive blocking their use by foreign nationals.GPT-5.6 SolOpenAI said its new AI product, called GPT-5.6 Sol, would be accessible only to customers approved by the Trump administration.Mythos 5Anthropic took offline two new AI models, known as Fable 5 and Mythos 5, just days after unveiling them to comply with a Trump directive blocking their use by foreign nationals.
Countries
Industries
Artificial intelligenceChatGPT maker OpenAI said Friday it is restricting the release of its new artificial intelligence model at the request of President Donald Trump’s administration, the latest in an unprecedented government vetting of AICybersecurityDonald Trump’s administration, the latest in an unprecedented government vetting of AI products for cybersecurity risks.