Claude Fable’s Relaunch Falls Flat Due to Nerfed Performance and Restrictive Limits
The highly anticipated relaunch of Anthropic’s powerful model, Claude Fable, has left many users underwhelmed. Initially available to all users, including those with a $100 Max subscription, the model comes with significant restrictions that have sparked disappointment among early adopters. While it appears that Anthropic hasn’t nerfed the model itself, its performance is being severely hampered by overly cautious safety guardrails.
Claude Fable’s relaunch was met with excitement when the Department of Commerce lifted the ban on the model in May. However, users were surprised to find that while Fable 5 is included in Max, Pro, and Team plans, it is heavily capped, limiting usage to up to 50% of weekly limits. After July 7, the model will transition entirely to a pay-to-play system via usage credits, adding an additional layer of complexity.
The real issue lies not with the model itself but with its degraded performance, or as the AI community calls it, “nerfed” performance. Users are reporting that Fable 5 feels weaker and is being routed through stricter safety systems more often than before, making it unreliable for everyday tasks. On Reddit, users have shared their experiences of encountering issues such as being forced to switch to Opus 4.8 mid-task or having the model completely fall back due to perceived safety risks.
The problem isn’t limited to Claude desktop; similar issues are also affecting Claude Code. Developers are finding that Fable’s increased sensitivity to prompts, project files, and security-related language is making it unusable for certain tasks. For instance, systems-level coding work involving languages like C, C++, Rust, or Win32 API references seems to trigger a fallback or block.
Anthropic has confirmed that its updated safeguards rely on a large “safety margin,” which could explain the subpar experience some users are seeing with Fable. While the company hasn’t acknowledged reports of false positives yet, it’s likely they are aware of the issue and will address it in a future update.
The Claude Fable relaunch debacle serves as a reminder that even the most powerful AI models can fall short when safety precautions take precedence over performance. As Anthropic continues to refine its safety protocols, users should be cautious of potential issues with their model choices. In the meantime, developers and security teams would do well to test every layer of their systems before attackers do.
Source: Bleeping Computer — 2026-07-03