Why the big comeback of Claude Fable 5 disappoints everyone

This week, Claude Fable 5 made its grand return in the Claude interface worldwide. After three weeks of suspension worldwide, Anthropic has reactivated the AI model for all its users.The Trump administration no longer considers artificial intelligence a threat to U.S. national security.

To regain favor with Washington, Anthropic says it has added new classifiers designed to block more cybersecurity-related tasks. The model is now available on Claude.ai, Claude Code, Claude Cowork, as well as on the Claude platform designed for businesses.

Also read: Anthropic unveils Claude Sonnet 5, its new default AI: here’s everything that changes

Pricing restrictions

Unfortunately, the highly anticipated return of Fable 5 disappointed many users. The main problem is that this return comes with serious limitations. As Anthropic explains, Fable 5 is included only up to 50% of the weekly usage quota for Pro, Max, and Team subscriptions, as well as certain Enterprise plans. Each subscription entitles the user to a certain amount of usage per week. A subscriber can use Fable 5, but this use cannot exceed half of their total weekly allowance. The other half of the quota remains reserved for standard models such as Claude Opus or Sonnet.

This is an important restriction because Fable 5 consumes about twice as many resources as Opus for equivalent use.In short, you use up your quota much faster with Fable 5 than with any other model. After July 7, the model will switch entirely to a pay-as-you-go system, operating through usage credits.

Slipping performance and overly strict guardrails

Beyond these pricing restrictions, many users are also criticizing Fable 5’s performance degradation. On Reddit, many users say the AI seems significantly less capable than the original version, which was banned in early June. Furthermore, Anthropic’s new guardrails appear to be triggered very often. When refusing to answer certain requests, Fable 5 regularly delegates tasks to Claude Opus 4.8. As one user put it, “it wasn’t the model that was banned”.

It is mainly developers who have been complaining about Fable 5’s performance since it was brought back online. Multiple reports mention blocks or automatic switches to another model whenever keywords such as “security,” “vulnerable,” or “insecure” are included in requests. Certain technical references also automatically trigger a switch to other AI models, which tends to frustrate users.

The problem of false positives

Everything points to Anthropic having tightened the screws a little too much in order to be allowed to reactivate Fable 5.In a blog post, Anthropic acknowledges that the new safeguards system is likely to generate “false positives”. These could result in “the blocking of more legitimate requests”, the startup admits. The company expects these false positives to be “frustrating for users, but we made this trade-off in order to make the model’s other features available to as many people as possible”. The company also specifies that this new classification system “has the drawback, however, of flagging benign requests more frequently during routine programming and debugging tasks”

I’ve heard a lot of questions about Fable’s availability on subscription plans.

While it will come off subscriptions after July 7th, we aim to restore Fable as a standard part of our subscriptions as soon as capacity allows, as we mentioned in our original blog post.

— Thariq (@trq212) July 2, 2026

Anthropic has not yet officially responded to criticism concerning the blocks and performance. However, Anthropic clarifies that the pricing restrictions related to Fable 5 will only be temporary. These measures, such as removing AI from subscriptions as of July 7, 2026, are only intended to give Anthropic a little time to secure enough computing power to run Fable 5 without limitations. The start-up plans to bring Fable 5 back into its service plans as soon as possible.

👉🏻 Follow the tech news in real time: add 01net to your sources on Google, and subscribe to our WhatsApp channel.

Source: Bleeping Computer