Anthropic Releases Claude Sonnet 4.6 Amid Pentagon Policy Tensions
According to the Last Week in AI podcast, Anthropic has released Claude Sonnet 4.6, the latest version of its AI model. The announcement comes alongside Google’s rollout of Gemini 3.1 Pro, marking significant updates from two major AI labs.
In a notable policy development, Anthropic CEO Dario Amodei addressed tensions with the Pentagon regarding AI work. According to the podcast, Amodei stated that the Pentagon’s threats “do not change our position” on AI collaboration, suggesting the company is maintaining its stance on defense-related AI projects despite apparent pressure.
The podcast also discussed “deep-thinking tokens,” though specific details about this technical development were not elaborated in the source material.
These developments highlight the ongoing evolution of frontier AI models and the growing policy debates surrounding AI companies’ relationships with government defense agencies. The simultaneous release of major model updates from both Anthropic and Google underscores the competitive pace of AI development among leading labs.