Anthropic's opus 5.2 and secret AI model signal rapid internal transformation

Anthropic has quietly deployed its new flagship model, Opus 5.2, within its Claude Code environment, with developers identifying it through network packet analysis. The model is already integrated into some real-world workflows, indicating an unannounced rollout. Simultaneously, an internal risk report revealed the company is utilizing an unpublished 'nuclear-level' Model 2, which has taken over most of its code writing tasks and is set to replace 85% of the research team. This marks a significant shift as Anthropic uses AI to replace its own staff rather than merely selling the technology externally. The dual development of Opus 5.2 and the internal Model 2 highlights a rapid acceleration in model iteration, with AI now deeply involved in code writing, surpassing prior expectations. The gradual release of Opus 5.2 contrasts with the already operational internal Model 2, which has begun reshaping the company's development structure. These developments suggest Anthropic is prioritizing internal efficiency and innovation, leveraging its own AI to redefine its research and development processes. The combination of these two initiatives underscores a broader trend of AI-driven transformation within the company, with implications for both its internal operations and external offerings.