Anthropic is preparing to launch a new AI model that combines language processing with reasoning abilities. According to an article by Matthias Bastian citing The Information, the model features variable resource allocation through a slider control. Users can adjust computing power for specific tasks, with the lowest setting operating as a standard language model. Early tests indicate strong performance in handling complex code bases. The company projects potential revenue of up to $34.5 billion by 2027, targeting API market dominance. The release is expected in the coming weeks, primarily serving enterprise customers. This development follows similar moves by OpenAI, as both companies shift away from traditional language model scaling due to diminishing returns.