Anthropic restricts access to Mythos and Fable AI models outside the US

Anthropic has restricted access to its latest AI models, Mythos 5 and Fable 5, to users within the US, a move that has reignited discussions around technological sovereignty and access to frontier AI systems.

The company unveiled Claude Fable 5 and Claude Mythos 5 as its newest generation of AI models. While Fable 5 will be available to users through Anthropic's products, access to Mythos 5 will be limited to a small group of organisations through Project Glasswing, the company's trusted-access programme.

Anthropic said the decision reflects the advanced capabilities of the models, particularly in areas such as cybersecurity. According to the company, Fable 5 is built on the same underlying technology as Mythos 5 but includes additional safeguards that restrict certain high-risk requests.

“We received the directive from the government today at 5:21pm (ET). The letter did not provide specific details of its national security concern. Our understanding is that the government believes it has become aware of a method of bypassing, or “jailbreaking” Fable 5. We reviewed a demonstration of this specific technique being used to identify a small number of previously known, minor vulnerabilities. These vulnerabilities all appear relatively simple, and we have found that other publicly-available models are able to discover them as well without requiring a bypass,” read the statement.

Mythos 5, meanwhile, will be made available only to approved organisations, including cyber defenders, critical infrastructure operators and select government-linked partners. The company said it intends to expand access gradually through a controlled programme rather than release the model broadly.

The company said that it disagrees with the government’s finding of a jailbreak. “However, we disagree that the finding of a narrow potential jailbreak should be cause for recalling a commercial model deployed to hundreds of millions of people. If this standard was applied across the industry, we believe it would essentially halt all new model deployments for all frontier model providers,” Anthropic noted.

Vembu says India must deepen AI research

The announcement drew a sharp response from Zoho chief scientist Sridhar Vembu, who said the move highlights the growing importance of technological self-reliance. "This is big: all access to Mythos and Fable AI models disabled for everyone outside America," Vembu wrote on X.

Calling technology the defining factor in national strength, he added, "Technology is the ultimate weapon. National sovereignty, national security, all of it is now about technology." Vembu argued that Indian organisations should increasingly adopt smaller open-source AI models, including those developed in India and China, rather than rely on foreign providers whose services may become unavailable.

"Why pay money to people who don't even want to sell to you?" he wrote. He also called for greater investment in domestic AI research and development, while acknowledging the difficulties involved in competing with frontier AI companies.

According to Vembu, training the most advanced models requires massive computing resources, access to high-end GPUs and investments running into tens of billions of dollars. He noted that access to such chips remains restricted and said India would need alternative approaches to remain competitive.

"Zoho has been pursuing alternative R&D approaches that are far, far less expensive but by its nature cutting edge R&D takes time and we are patient. I am confident we will get there," he said.

Vembu ended his post with a broader warning about the changing nature of global technology markets. "Any remaining people in India who have delusions about globalization should wake up now."