Anthropic Releases Claude Fable 5, Its First Public Mythos-Class AI Model
6 verified2 unconfirmed
Anthropic has launched Claude Fable 5, the first broadly available version of its powerful Mythos class of AI models. The company stated the model excels at software engineering, knowledge work, and vision tasks, with performance growing stronger on longer and more complex assignments. Fable 5 includes new safety guardrails that block responses in high-risk areas such as cybersecurity, biology, chemistry, and distillation, falling back to the Claude Opus 4.8 model when triggered. Pricing for both Fable 5 and the companion Mythos 5 is set at $10 per million input tokens and $50 per million output tokens, double the cost of the previous flagship Opus 4.8. Anthropic reported that in testing, at least 95 percent of Fable sessions ran entirely on the model’s own responses without requiring a fallback. The company is also deploying Mythos 5 to organizations that were previously approved to access the advanced model.
What’s verified
Anthropic released Claude Fable 5, the first publicly available version of its Mythos-class AI model.
The model is described as excelling in software engineering, knowledge work, and vision.
Safety guardrails block responses in high-risk domains (cybersecurity, biology, chemistry, distillation) and cause the model to fall back to Claude Opus 4.8.
Pricing is $10 per million input tokens and $50 per million output tokens, double the rate of Opus 4.8.
Anthropic stated that 95% of Fable sessions run entirely on Fable’s own responses in testing.
The company is also releasing Claude Mythos 5 for organizations already approved to access the advanced model.
Not yet confirmed
One source detailed that Fable 5 access through subscriptions will be included at no extra cost until June 22, then pulled and replaced with usage credits, with plans to restore it as a standard feature later. This source also reported a 30-day data-retention policy for all traffic and an external bug bounty that found no universal jailbreaks in over 1,000 hours of testing.
Another source reported that Anthropic did not on the record explain the model numbering or the relation to the earlier Claude Mythos Preview. This source also noted that Mythos 5 access is currently limited to organizations in a private initiative called Project Glasswing.
Key figures
Anthropic (company), Claude Fable 5 (AI model), Mythos 5 (AI model), Claude Opus 4.8 (AI model)
Sources: TechCrunch, The Verge
