Anthropic, the company behind the Claude chatbot, has released Fable—a powerful new AI model described by early users as a step change in capability. Built as a publicly safe variant of Mythos, which the company deemed too dangerous to release two months ago, Fable incorporates stringent guardrails to prevent high-risk misuse.
For now, Fable is accessible to subscribers on Pro, Max, Team, or Enterprise plans through June 22. After that date, Anthropic will restrict access solely to API users, who must pay per token. The limited window reflects a growing reality: advanced AI models come with steep operational costs, and tokens are far from free. Anthropic says the temporary consumer availability is due to capacity constraints.
The move signals a broader trend in the AI industry: powerful models are increasingly gated by cost and safety concerns. While Fable's capabilities dazzle, its rapid shift to an expensive API model may limit adoption among smaller businesses and individual developers. Enterprise customers, accustomed to per-token pricing, may find the model more accessible.
For competitors, Fable raises the bar. OpenAI and Google have released their own frontier models, but Anthropic's emphasis on safety—and its willingness to withhold a model entirely—sets it apart. The question now is whether businesses will pay the premium for Fable's performance or seek cheaper alternatives.
Anthropic has not disclosed pricing details for the API tier, nor has it commented on when additional capacity might allow broader consumer access. The company's focus remains on balancing capability with responsibility, even as costs climb.