ANTHROPIC TO RELEASE MYTHOS-CLASS AI MODELS WIDELY

Anthropic is moving forward with a broad release of AI models comparable to Mythos, technology the company had deemed too risky for public access. The San Francisco-based AI firm says the new models will be available to all customers "in the coming weeks" after implementing enhanced safeguards. The announcement comes as Anthropic released Claude Opus 4.8 on Thursday, an upgraded version of its flagship model featuring improved coding and knowledge work capabilities. The new version maintains pricing at the same level as its predecessor. Mythos-class models represent a significant capability tier in Anthropic's product lineup, with cybersecurity features on par with the original Mythos system. The company's previous hesitation to release these models reflected broader industry caution around deploying advanced AI systems without adequate safety measures in place. Anthropric's decision to proceed with the wider rollout signals confidence in its safety infrastructure. The development of stronger safeguards—specific details of which remain undisclosed—apparently addressed the company's initial concerns about unrestricted access. The release strategy aligns with Anthropic's competitive positioning against other major AI labs, including OpenAI and Google DeepMind. As the AI market intensifies, companies face pressure to make powerful models accessible while managing associated risks. Anthropic has emphasized responsible AI development throughout its operations, including publishing research on AI safety and implementing various protective measures across its product line. The Mythos release represents a test of whether those safeguards can scale to broader user bases. The timing of the announcement—during an accelerating period of AI model releases and upgrades—underscores the rapid pace of development in the sector. Anthropic's willingness to move from restricted to public access suggests the company believes its risk mitigation strategies have matured sufficiently for general deployment.