Anthropic announced on Tuesday the launch of Claude Fable 5, a new AI model the company says is the most powerful it has ever made available to the public.
According to Anthropic, Fable 5 delivers exceptional performance in software engineering, knowledge work, and vision tasks. The company said that the model's advantage over competing systems becomes more pronounced as tasks grow longer and more complex.
First Mythos-Class Release
Fable 5 marks the first broad release from Anthropic's Mythos family of AI models. The company had previously stated that this model class was so capable at cybersecurity tasks that it posed too great a risk for public release. Now, Anthropic says new safety measures have made the public launch possible.
The safeguards work by blocking the model's responses in specific high-risk areas. When those blocks are triggered, the system falls back to Claude Opus 4.8, a model Anthropic praised for its honesty when it launched last month. During testing, the company said that 95 percent of Fable sessions ran entirely on Fable responses, without needing to fall back to Opus 4.8.
Anthropic singled out cybersecurity and biology as two domains where the safeguards may block responses. Both areas are widely considered sensitive topics for advanced AI systems.
Claude Mythos 5 Also Released
Stay updated
Get the day's AI and automation news in your inbox. No spam, unsubscribe anytime.
Alongside Fable 5, Anthropic is releasing Claude Mythos 5. In a blog post, the company described Mythos 5 as the same underlying model as Fable 5 but with the safeguards lifted in some areas. Access to Mythos 5 appears limited for now to organizations that are part of Anthropic's Project Glasswing initiative, a private and not entirely watertight program that steadily expands the number of trusted users. Those users can upgrade to Mythos 5, and Anthropic said it plans to expand access over time through a more systematic trusted-access program.
Anthropic did not respond on the record to The Verge's request for comment explaining how either model relates to Claude Mythos Preview or why the models are numbered "5" when there do not appear to be any previously released Mythos or Fable models.
Pricing and Access
Pricing for both models is significantly higher than Anthropic's former flagship model. Fable 5 and Mythos 5 cost double the rates of Claude Opus 4.8, though they are half the price of Mythos Preview. The cost is $10 per million input tokens and $50 per million output tokens, Anthropic said.
The release represents a shift in Anthropic's stance on releasing highly capable AI models. The company had previously argued that such powerful systems could be too dangerous for public use, but now believes new safeguards are sufficient to mitigate the risks.