Anthropic’s new AI model is about to get expensive

Fable shifts – Anthropic’s new Fable model—built as the publicly safe version of Mythos—will move behind API pricing after June 22. Monthly subscribers on Pro, Max, Team, or Enterprise plans can use it only until then, with access later shifted to per-token charges, as capac

On June 22, Anthropic will pull Fable out of most people’s subscription plans—right as the model is being described as a step change in capability.

Built by the company behind the Claude chatbot, Fable is the publicly safe version of Mythos. Two months earlier, Anthropic had deemed Mythos too dangerous to release. To make Fable available to a wider audience. Anthropic has put it under stringent guardrails meant to prevent the cutting-edge model from being used for high-risk applications.

People who have used Fable describe it as powerful enough to handle complicated tasks with ease. That’s exactly the kind of pitch that drives adoption—and, for subscribers, it comes with a deadline.

For now, Fable is accessible to monthly Anthropic subscribers on Pro, Max, Team, or Enterprise plans. But that access lasts only until June 22. After that, Anthropic says it will place Fable solely behind API access, where users pay on a per-token basis. The company is also signaling that this shift may not stay permanent, but the timing is uncertain.

Anthropic says the reason for the limits is capacity. “As enough capacity comes online. we aim to make it a standard part of those [subscription] limits again. ” wrote Amol Avasare. Anthropic’s head of growth. on X on June 9. 2026. He added: “We’re sprinting as hard as we can at this. ” while also saying the company couldn’t make any promises on timing.

Once the subscriber window closes, customers will still be able to use Fable 5 via Extra Usage, but it will not be included in subscription limits.

The immediate impact is straightforward: even users who never approached their monthly caps may soon feel their budgets tighten. The broader impact is more uncomfortable. Since the release of ChatGPT in November 2022. consumer AI has largely followed a familiar script—pay $20 or more a month. then use the smartest systems as much as the company allows. Many users never reached their limits, while heavy users often benefited from effectively subsidized access.

Now. as agentic AI replaces chatbots as the newest way to interact with AI systems. the economics of running these models are getting harder to ignore. Inference—the cost of delivering the service—has shot up. As systems become more capable. people increasingly ask them to do longer. messier. more valuable work. even as the hardware needed to serve the models keeps getting more expensive.

That’s pushing the industry quietly away from “all you can eat” toward “eat what you can afford.” The shift is already visible elsewhere: last month. Google shifted Gemini toward compute-based limits; Microsoft’s GitHub recently cut back its usage limits; and at the end of May. OpenAI switched off its usage-limit multiplier after a monthslong trial.

Taken together. Anthropic’s June 22 deadline fits a wider pattern—AI access is being re-priced from flat-fee convenience to consumption-based reality. Fable may still be the model people want to test while it’s included in subscription budgets. but after that. the question changes from what it can do to what it will cost each time it’s asked to do more.

So for subscribers, the message is simple and time-sensitive: try it now, because after June 22 the bill won’t be capped the same way.

Anthropic Fable Mythos Claude AI pricing API access per-token pricing capacity constraints Amol Avasare inference costs agentic AI ChatGPT Gemini GitHub usage limits OpenAI usage-limit multiplier