Claude Fable 5 (www.anthropic.com)

Claude has launched two advanced AI models: Claude Fable 5 and Claude Mythos 5. Fable 5 is a Mythos-class model designed for general use, showcasing unprecedented capabilities across multiple domains, including software engineering, knowledge work, vision tasks, and scientific research. It excels particularly in complex and lengthy tasks, outperforming previous models by a significant margin. However, acknowledging the potential risks associated with such powerful capabilities, Claude has integrated conservative safeguards that redirect sensitive queries to their less powerful counterpart, Claude Opus 4.8, triggering on average in less than 5% of interactions. Mythos 5, on the other hand, offers lifted safeguards for select cyberdefenders and infrastructure providers, marking it as the most powerful model in cybersecurity. Early user feedback highlights Fable 5’s superior performance in tasks ranging from autonomous coding to complex analytical reasoning, significantly accelerating engineering processes and enhancing judgment in financial contexts. Both models are priced affordably, aimed at broadening access to cutting-edge AI capabilities while maintaining safety protocols. As these models demonstrate their impact in crucial fields like drug design and molecular biology, Claude emphasizes that advancements come with an ongoing commitment to refining safety measures to mitigate misuse risks.