Anthropic Plans Wide Release of Mythos-Level AI Models in Weeks

Anthropic plans to widely release new artificial intelligence models in the coming weeks with cybersecurity capabilities comparable to Mythos, a technology the company previously said was too dangerous to make available to the general public.

The company said Thursday that it has made “swift progress” in developing “stronger safety safeguards” that would allow it to release Mythos-level AI models to all customers. The company made the announcement alongside the release of a new model, called Opus 4.8, that’s meant to be better at carrying out coding tasks on behalf of users.

Anthropic has said Mythos is capable of identifying and exploiting vulnerabilities “in every major operating system and every major web browser when directed by a user to do so.” As a result, the company initially limited its availability to a handful of large tech and Wall Street companies through an initiative known as Project Glasswing in April, amid broader global alarm about the new threats Mythos could pose to critical systems.

On Friday, Anthropic said it planned to work with the US and allied governments to expand Project Glasswing to more partners. The company also said it intended to make models similar to Mythos more widely available, but did not provide a timeline for doing so.

The Claude maker is locked in a heated competition with OpenAI to deploy better artificial intelligence models and convince more business customers to pay for them. In recent months, Anthropic has seen strong momentum for its AI coding offerings as well as growing traction with consumers amid a standoff with the Pentagon over AI safeguards.

Anthropic is nearing the close of a new funding round at a more than $900 billion valuation, vaulting it ahead of OpenAI’s most recent valuation. Anthropic and OpenAI are also both discussing going public as soon as this fall, Bloomberg News has reported.

Photo: Photographer: Gabby Jones/Bloomberg