Anthropic has made its most powerful artificial intelligence model, Claude Fable 5, publicly available for the first time. This is the first public version of the Mythos model, which excels in software development, analytics, and visual data processing.
Anthropic has made its most powerful artificial intelligence model, Claude Fable 5, publicly available for the first time. This is the first public version of the Mythos model, which excels in software development, analytics, and visual data processing.
According to the company, in high-risk areas such as cybersecurity, biology, or chemistry, Fable 5 will automatically block responses and switch to the weaker but proven Claude Opus 4.8 model. This is due to concerns about the technology's potential use by attackers.
Now, anyone can access the technology via the Claude API and Enterprise plans. Until June 22, Fable 5 will be available at no additional cost for Pro, Max, Team, and Enterprise subscriptions. However, starting June 23, the company will change the rules - using the model will require separate credits, although Anthropic promises to return it as a standard feature later.
“We ran an internal bug bounty program, which failed to find any universal jailbreaks after over 1,000 hours of testing. We also engaged external teams, who also failed to find critical vulnerabilities,” Anthropic said.
At the same time, the company is introducing an unprecedented rule - all users of Fable 5 and the updated Mythos 5 will be required to retain traffic for 30 days. This applies even to those companies that previously had zero-data retention agreements. Anthropic assures that the information will not be used to train AI, but will serve exclusively to protect against new types of attacks and jailbreaks.
The cost of using both new models (Fable 5 and Mythos 5) is $10 per million incoming tokens and $50 per million outgoing tokens. This is twice the cost of Opus 4.8. The high price can be a significant deterrent for many businesses that are already facing cost overruns in implementing AI.
Previous versions of Mythos, launched in April, were only available to a limited number of partners, mostly organizations managing critical infrastructure, due to cybersecurity risks.
It was previously reported that Anthropic's specialized Claude Mythos artificial intelligence model helped a team of researchers uncover a critical vulnerability in Apple systems . The version of the AI, which the company currently provides only to select partners, allowed specialists to bypass macOS kernel memory protection on the latest M5 processors.
