Fable 5 is back, and it hands your code to Opus 4.8

It won't really write your code

Anthropic is redeploying Fable 5 with a new set of classifiers aimed at cybersecurity misuse. The side effect: routine coding and debugging will fall back to Opus 4.8, benching the flagship model for the exact task most developers wanted it for.

Anthropic says it will tune the classifiers over the coming weeks to cut false positives. Until then, counting on Fable to write your code means counting on Opus 4.8.

One week before the meter starts

On Pro, Max, Team and some Enterprise plans, Fable 5 is included for up to 50% of your weekly limits, and only through July 7. After that you are on pay-as-you-go credits.

That is why the community sums up the return in one line: you have a week to use Fable 5 without going bankrupt. It is a short window, not a free comeback.

You can get downgraded mid-chat

Trip a policy classifier with your prompt and you can be routed to Opus 4.8 without really choosing to. Hence the loudest complaint: if it will not code and sometimes will not even answer as Fable, what is it for — chatting?

How to use the window

The savvy play: stop using Fable as your code writer. Use it as a planner — let it chew on the hard architectural problems and propose solutions — and hand the actual implementation to cheaper or less-restricted models.

In practice: plan with Fable, delegate the typing to another model, and save the expensive calls for what actually needs judgment.

The context: government, classifiers and a consensus framework

The return follows a series of conversations with the US government. Anthropic also says it is drafting a consensus framework with Amazon, Microsoft, Google and other partners to assess the severity of AI jailbreaks.

It will also re-enable access via AWS, Google Cloud and Microsoft Foundry as fast as it can. So it may loosen up. For now, the window is short and the model is fenced off from its best use case.