Anthropic has redeployed Claude Fable 5 globally after a three-week hiatus, implementing a new safety classifier that automatically reroutes certain cybersecurity-related tasks to the less capable Opus 4.8 model. This change, while intended to enhance security, has led to increased false positives for routine coding and debugging tasks, frustrating some users who perceive it as a downgrade or a "bait-and-switch." Anthropic is also collaborating with major tech partners and the US government to develop a framework for assessing and responding to AI jailbreaks. AI
IMPACT This shift may lead to more false positives for developers, potentially impacting workflows and increasing reliance on less capable models for certain tasks.
RANK_REASON Frontier-lab model release with system card and new behavior.
AI-generated summary · Google Gemini · from 9 sources. How we write summaries →