Anthropic is set to release a new hybrid AI model that combines reasoning capabilities with traditional language model functions, The Information reported on Friday, citing people who have used the new model.
According to the report, the new model would allow developers to control how much computational power is used per query. Unlike OpenAI's fixed "low," "medium," and "high" reasoning settings, Anthropic's model features a sliding scale for greater cost and speed flexibility.
The model excels in coding tasks, particularly in understanding complex codebases and generating accurate lines of code. Moreover, it reportedly outperforms OpenAI's top reasoning model, o3-mini, when given maximum processing time.
