Anthropic is making its most advanced AI model available to the general public for the first time, though it comes with built-in safeguards. On Tuesday, the AI company introduced Claude Fable 5, the first publicly accessible version of its Mythos model. According to Anthropic, Fable 5 performs exceptionally well in software engineering, knowledge work, and vision tasks, but it has strict safety restrictions. In high-risk domains such as cybersecurity, biology, chemistry, and distillation, the model refuses to answer and instead defaults to Claude Opus 4.8.

First unveiled as a preview in April, Mythos was originally restricted to a small group of partners due to cybersecurity concerns. Last week, Anthropic broadened access to hundreds of organizations across 15 countries, once again prioritizing those managing critical infrastructure. Now, a version of this technology is available to anyone through Anthropic’s Claude API and consumption-based Enterprise plans. Access through subscriptions will be introduced gradually: until June 22, Fable 5 will be included in Pro, Max, Team, and seat-based Enterprise plans at no additional cost. Starting June 23, Anthropic will remove Fable 5 from those plans, requiring usage credits going forward, with plans to reinstate it as a standard subscription feature as soon as possible.

Anthropic is also rolling out a new version of Mythos, called Mythos 5, to organizations that have already been approved for the advanced model. The launch of Fable coincides with Anthropic’s preparations to enter public markets, alongside OpenAI and Elon Musk’s SpaceX. It also follows the AI firm’s call for major global AI labs to establish a coordinated emergency brake on frontier AI development. Anthropic has warned that systems are advancing so quickly that they may soon achieve recursive self-improvement (RSI), allowing them to autonomously enhance themselves without human intervention.

Concerned about what a Mythos-class model could do in the wrong hands, Anthropic says it rigorously tested its classifiers with jailbreak attempts before releasing Fable 5. “Internally, we ran an external bug bounty that produced no universal jailbreaks in over 1,000 hours of testing. We then worked with external red-teaming orgs which also failed to find universal jailbreaks,” the company stated.

That said, novel attacks could still emerge. As a result, with the launch of Fable 5 and Mythos 5, Anthropic announced it will require a 30-day retention on all traffic, even if enterprises previously had zero-retention agreements. The company says it will not use the data for training and will only use it to “defend against complex and novel attacks, including new jailbreaks,” and “identify and reduce false positives.” This policy could set an industry precedent, where access to increasingly powerful models comes with mandatory data-retention policies framed as a safety measure.

For those who continue using the model, not every question will receive a Fable 5 answer. Anthropic says the instances where Fable must defer to Opus 4.8 are rare, with early data showing that at least 95% of Fable sessions run entirely on the model’s own responses. In third-party testing, analytics company Hex stated that Fable was the first to achieve a 90% score on its core analytics benchmark of complex, long-running analytical tasks. “On the hardest questions, it shows strong judgement and attention to nuance,” Hex noted. Vibe-coding platform Base44 observed that Fable is better at “one-shotting full apps” and has excellent tool-calling capabilities. AI-powered workspace and agent platform Genspark reported that Fable outperformed every other model in its evaluations and performed significantly better on tasks like UI design and game coding.

Pricing for both Fable 5 and Mythos 5 is set at $10 per million input tokens and $50 per million output tokens, double the cost of Opus 4.8. That price alone might discourage widespread use. Many enterprises are becoming increasingly critical of AI costs after seeing their bills or blowing through their yearly AI budgets early. Advanced models like Opus 4.8 can exacerbate these issues, with advanced reasoning skills that can split a single request into multiple tasks. Anthropic says it expects demand for Fable 5 to be very high and difficult to predict. And indeed, some, like shopping rewards platform Rakuten, might see the upside as worth the price point. “At the highest effort, Fable reflects on and validates its own work,” Rakuten said in a statement. “For us, that’s what makes highly autonomous operations possible - the extra thinking pays for itself.”