Anthropic Offers Mythos Upgrade for Cyber Partners and a ‘Safe’ Version for the Rest of You (www.wired.com)

Anthropic has unveiled two new AI models, Claude Fable 5 and Claude Mythos 5, with enhanced capabilities compared to the earlier Mythos Preview model. While Claude Mythos 5 is being released to a select group of tech industry partners and biology researchers—alongside collaboration with the US government—Claude Fable 5 is available to the public but comes equipped with significant restrictions. These safeguards are designed to prevent the AI from addressing sensitive inquiries related to cybersecurity and other critical fields, redirecting such queries to the older model, Claude Opus 4.8, to mitigate potential misuse. This release is significant for the AI/ML community as it underscores the balancing act between advancing AI capabilities and ensuring safety against exploitation. Anthropic's efforts highlight ongoing cybersecurity concerns as AI models gain the ability to identify and exploit software vulnerabilities. The company admits that while the current safeguards are designed to be cautious, further refinements will be made to enhance accuracy over time. With anticipated future expansions to access and competitive developments in AI, both Anthropic and its rivals are under pressure to securely deliver advanced capabilities that maintain user safety in a rapidly evolving technological landscape.