U.S. allows Anthropic to restore limited access to its Mythos 5 AI model
The Facts
- The U.S. government authorized Anthropic on Friday to restore access to its Mythos 5 AI model for a limited set of trusted or approved organizations.
- The partial restoration came about two weeks after the government ordered Anthropic to suspend access to some advanced models because of national security concerns.
- Anthropic had disabled access to both Mythos 5 and Fable 5 after the government directive.
- More than 100 companies and institutions are expected to receive access to Mythos 5 under the new arrangement.
- The approved users include U.S. companies and some government agencies or institutions, meaning access is being restored selectively rather than broadly.
- Fable 5 remains restricted for now, and discussions were still underway about restoring access to that model and expanding Mythos 5 access further.
- The dispute has become part of a broader U.S. effort to more tightly oversee releases of advanced AI models from Anthropic and rival OpenAI.
Context
Why did the U.S. restrict Anthropic's models in the first place?
The government said the restrictions were tied to national security concerns, including fears that advanced models could be accessed or misused by foreign nationals or by military intelligence users in countries such as China and Russia uol.com.br,News18,India Today.
Who can use Mythos 5 now?
Access is being restored only to a limited group of approved or trusted organizations, with reports saying more than 100 companies and institutions are included and that some federal agencies are among the eligible users uol.com.br,POLITICO,CNBC.
Where Left and Right agree, and where they split
- Where Left and Right agree
- Selective restoration of Mythos 5 while Fable 5 stays restricted shows advanced AI access is now being governed through targeted national-security oversight.
- Where Left and Right split
- Concentrated government power over who gets advanced AI access, or proving that such power can be exercised with disciplined restraint.
How left and right read it
Left says
The striking fact here is not just that access to Mythos 5 is back for approved users, but that it is being restored selectively while Fable 5 remains restricted and further expansion is still under discussion. That matters because it shows advanced AI is increasingly governed through concentrated approval power, with access shaped by national security judgments and a broader push to tighten oversight of powerful models.
Right says
What stands out here is institutional restraint: the government eased access to Mythos 5 for approved users while keeping Fable 5 restricted and continuing discussions about any broader expansion. That matters because tighter oversight of advanced AI only holds public legitimacy if it is disciplined, selective, and tied to national-security concerns rather than treated as a blank check.
The receipts — 94 sources
Wire services (2)
Independent coverage (92)
India TodayNo Mythos for Indians: US govt tells Anthropic to give Mytho...
Il Fatto QuotidianoAnthropic Claude Mythos 5, il governo Usa revoca il blocco: ...
RTOpenAI restricts ChatGPT rollout at Washington's request -- ...
Economic TimesInside Claude Mythos: Why Anthropic held back its most advan...
CumhuriyetOpenAI, yeni yapay zeka modellerini sınırlı ön izlemeye açıy...
Economic TimesOpenAI unveils GPT-5.6 AI models with phased rollout and str...
Deutsche WelleUS-Regierung verstärkt Einfluss auf KI-Entwickler
Deutsche WelleUS allows partial release of Anthropic's Mythos AI model
India TodayOpenAI announces Sol, Terra and Luna AI models, US govt bloc...
LaVanguardiaEE.UU. vuelve a autorizar IA más potente de Anthropic para u...
The HinduU.S. allows limited access to Anthropic's Mythos AI model
The Times of IndiaAnthropic says US government has restored Mythos 5 access, b...
En Son HaberABD, yapay zeka şirketi Anthropic'in 'Mythos' modeline yönel...
NDTVAnthropic Receives Approval To Release Claude Mythos AI Mode...
HaberlerABD, yapay zeka şirketi Anthropic'in "Mythos" modeline yönel...
mintOpenAI names Prabhjeet Singh India MD, launches GPT-5.6 Sol ...
mintTrump administration rolls back part of Anthropic model ban ...
rtl.deUS-Regierung lässt Zugang zu ChatGPT-Version einschränken
XatakaMythos vuelve solo para un selecto club de empresas estadoun...
El Financiero¿Por qué Trump 'restringió' la nueva versión de Chat GPT 5? ...
N-tvChatGPT 5.6 von US-Regierung eingeschränkt - nur US-Firmen h...
TimesNowUS Government Clears Limited Rollout Of Anthropic's Claude M...
Yahoo!Gobierno de EE.UU. autoriza a Anthropic publicar de forma li...
CNN EspañolGobierno de EE.UU. autoriza a Anthropic publicar de forma li...
cnbctv18.comOpenAI unveils GPT 5.6 but limits access to government appro...
NDTVOpenAI's New AI Was So Powerful, It Had To Brief The US Gove...
cnbctv18.comAnthropic's Claude Mythos 5 returns as US clears access for ...
Business StandardAnthropic's Mythos 5 AI model gets approval from Trump admin...
Los Angeles TimesAnthropic moves toward deal with U.S. to lift curbs on AI mo...
The Indian ExpressAnthropic resumes limited Mythos 5 rollout after US eases re...
NBC NewsU.S. government gives Anthropic green light for limited re-r...
Spiegel OnlineUS-Regierung lässt Zugang zu neuem OpenAI-Modell beschränken
The Times of IndiaUS approves limited rollout of Anthropic's Mythos 5 AI model
La Repubblica.itAnthropic, gli Stati Uniti revocano il blocco su due modelli...
The Straits TimesUS allows Anthropic to release Mythos to 'trusted partners'
The HillFederal government permits release of Anthropic's Mythos mod...
mintTrump administration clears Anthropic to roll out Claude Myt...
TerraEUA suspendem bloqueio e autorizam acesso limitado ao modelo...
The Straits TimesOpenAI defers public roll-out of GPT‑5.6 as US seeks early a...
EstadãoEUA suspendem bloqueio e autorizam acesso limitado ao modelo...
India TodayUS allows Anthropic to release Claude Mythos 5 to trusted pa...
The Financial ExpressUS allows Anthropic to restore Mythos 5 for 'trusted partner...
MoneyControlAnthropic's Mythos 5 AI model cleared by US for wider use- M...
News18US Approves Limited Release Of Anthropic AI Model After Cybe...
AxiosAnthropic's Mythos is coming back for a select group of enti...
Economic TimesUS allows Anthropic to release Mythos to 'trusted partners'
The VergeAnthropic's Mythos 5 is back
ANSA.it++ Usa revocano il blocco su due modelli IA di Anthropic ++ ...
CNN InternationalUS government allows Anthropic limited release of AI model t...
The New York TimesU.S. Loosens Restrictions on Anthropic's Mythos A.I. Model
The Wall Street JournalTrump Administration Rolls Back Part of Anthropic Model Ban
POLITICOTrump administration partially lifts Anthropic's AI export b...
TerraGoverno dos EUA autoriza Anthropic a liberar Mythos para alg...
Semana.com Últimas Noticias de Colombia y el MundoInesperada petición de Trump: Nuevo modelo de tecnología que...
CNBCTrump admin allows Anthropic to release Mythos AI model to s...
PerfilOpenAI lanzó una versión preliminar de GPT-5.6 con acceso li...
Washington PostOpinion | Feds deciding who gets GPT access reflects a misre...
ZEIT ONLINEOpenAI: OpenAI präsentiert neues KI-Modell für ausgewählte K...
The HillOpenAI slow rolls new model amid Trump administration concer...
iXBT.comGPT-5.6 уже вышла, но не для всех: США решили контролировать...
GoodreturnsOpenAI GPT-5.6 Sol release restricted after US government cy...
Yahoo!La nueva versión de ChatGPT en pausa: Gobierno de EE.UU. tem...
Blick.chOpenAI lance un nouveau modèle pour rivaliser avec Anthropic
Yahoo! FinanceOpenAI restricts limited release of new model to US only
EL MUNDOAsí es GPT-5.6, la IA más avanzada que, de momento, no podrá...
FortuneOpenAI agrees to stagger rollout of its most powerful model ...
Milenio.comOpenAI abre sus nuevos modelos de GPT a socios aprobados por...
Jornal Estado de Minas | Not�cias OnlineOpenAI lança modelo de IA apenas nos EUA, a pedido de Trump
HaberlerOpenAI, yeni yapay zeka modellerini ABD hükümetinin talebiyl...
GoodreturnsOpenAI GPT-5.6 Sol release restricted after US cybersecurity...
La RazónGPT-5.6 Sol: los modelos de nueva generación de OpenAI se pr...
Фонтанка.руOpenAI выпустила три новых версии GPT-5.6. Но пока не для вс...
uol.com.brIAgora?: Você não pode usar (ainda) nova IA poderosa do Chat...
NDTVOpenAI Launches GPT-5.6 Series in US-Only Preview At Governm...
Yahoo! FinanceOpenAI limits its latest ChatGPT product to Trump-approved c...
El UniversalOpenAI lanza GPT 5.6, su nuevo modelo de inteligencia artifi...
La Presse.caAccès limité | OpenAI lance son nouveau modèle GPT 5.6
infobaeOpenAI confirma un acceso limitado a sus tres nuevos modelos...
La RazónLa Casa Blanca quiere controlar quién puede usar GPT-5.6, el...
Business InsiderOpenAI says access to its new GPT-5.6 model is limited at th...
The HillOpenAI slow rolls new model release at 'request' of governme...
Washington PostU.S. government will decide who gets to use latest upgrade t...
AmbitoOpenAI retrasa el lanzamiento de GPT-5.6 por presiones del g...
EL MUNDOEl gobierno de EEUU decidirá quién accede al próximo modelo ...
nsctotal.com.brCasa Branca pede que OpenAI desacelere novo ChatGPT por teme...
Olhar Digital - O futuro passa primeiro aquiPor que a OpenAI não vai liberar o GPT-5.6 para todo mundo?
infobaeOpenAI retrasa el lanzamiento de GPT-5.6: Gobierno de Estado...
Le Figaro.frChatGPT : l'administration Trump demande à OpenAI de retarde...
XatakaEEUU aprieta las tuercas a OpenAI: debe restringir el lanzam...
uol.com.brTrump pede que OpenAI atrase lançamentos de IA por questões ...
ForbesOpenAI Rolls Out Powerful GPT-5.6 Models To Limited Users Ve...
newsORF.atOpenAI präsentiert neues KI-Modell
Facts first. Then every angle.
The day’s biggest stories in one short brief — the facts everyone agrees on, then the competing values behind the headlines. Free in your inbox.