Anthropic spent months saying that Mythos was too dangerous to publish; then released a public version called Fable 5, which it warns "comes with risks."

For the past few months, Anthropic’s most powerful AI model has largely existed as a warning about the dangers of AI. The company declined to go public and repeatedly described its Mythos-class systems as capable enough to raise serious concerns about cybersecurity, biological research (think nasty pathogens), and the accelerated pace of AI development.

The message was clear: this was a technology that required extraordinary safeguards before it could be launched.

Now Anthropic has done something surprising. It has released a public version of that same technology, in Claude’s top-tier subscription plans, if only for a limited period.

This is what Fable 5 does well

According to Anthropic, Fable 5 is designed to do much more than you might be using AI right now, like answering questions or helping you compose an email. The company says it’s designed to handle the kind of work that takes place over hours, days, or even longer, so we’re talking about everything from software engineering projects and in-depth research to complex AI agent workflows.

The big idea is that Fable 5 can perform a task for much longer than previous models. Instead of responding to a single prompt and waiting for the next instruction, you’re supposed to be able to solve multi-step problems, track context across long projects, and progress with a greater degree of autonomy.

You can use Fable 5 through the Claude chatbot, but many of its biggest improvements seem to be aimed at advanced users, researchers and developers. The goal seems to be to move AI beyond a conversational assistant and into something that feels like a true digital collaborator.

Why cybersecurity is at the center of the debate

Much of the concern surrounding Mythos relates to cybersecurity, with Anthropic stating that the model has “the strongest cybersecurity capabilities of any model in the world.” While those capabilities could help security researchers identify and patch vulnerabilities, they could also be used to discover and exploit the same software vulnerabilities if placed in the wrong hands.

Anthropic recognizes that “launching a model with this capability carries risk” and is attempting to address that risk through a unique set of safeguards. The model is combined with separate artificial intelligence systems known as “classifiers” that monitor requests for signs of misuse. When those classifiers detect a request related to an advanced cybersecurity activity, the response is handled by Claude Opus 4.8 instead of Fable 5, reducing the risk.

In fact, Anthropic allows public access to Mythos-class capabilities while also placing some of the most sensitive areas behind an additional layer of protection. That sounds great, as long as it works.

A positive reception

Initial reaction from users has been positive, although they have noted how quickly Fable 5 burns tokens.

“Fable 5 is incredibly good, but watch your usage, I was burning 2% per minute on 20x,” said one Reddit user, who also notes, “I’m on the Max 20x plan and during a heavier session I saw my usage increase by about 2% per minute. Not hourly. Per minute.”

Another user takes issue with Fable 5’s refusal to address cybersecurity, calling it a “preview of AI inequality,” noting that “the public gets the ‘safe’ version. Trusted institutions get the dangerous/useful version.”

If you want to use Fable 5, there are some quirks to its availability that are worth knowing about.

Starting today, Fable 5 is available on seat-based Claude Pro, Max, Team, and Enterprise plans at no additional cost. However, Anthropic says that starting June 23, access will temporarily move to a usage credit system.

That means subscribers who want to continue using Fable 5 beyond that date may need to purchase additional usage credits depending on how much they use the model.

Anthropic says the change is aimed at managing demand, rather than creating a permanent paywall. “After this point, when sufficient capacity allows, our goal is to restore Fable 5 as a standard part of subscription plans,” the company states. “We intend to do this as quickly as we can.”

While early users seem satisfied with Fable 5’s capabilities, its availability highlights how advanced AI systems are now becoming and may point the way to a future where the public will no longer have full access to the most advanced models.

Google logo on black background next to text that says