In July, Bloomberg reported that OpenAI had defined five stages of AI innovation, with the first being chatbots like GPT-4o, and the second being "reasoners," capable of human-level problem solving. It's an open secret in Silicon Valley that OpenAI is currently deep in development on its own reasoning model, named Strawberry. Such a model could be capable of reasoning through problems in a multi-step process, making them better equipped to deal with challenges that current models struggle with, such as solving complex math problems. The Information previously reported that Strawberry could be released as soon as this fall.
You are viewing a single comment's thread from: