OpenAI on Thursday introduced o1, its latest large family of language models, which it claims can emulate complex reasoning.

The O1 model range – which currently consists of the o1 preview and the o1 mini – uses “chain of thought” technologies.

In a 2022 paper, Google researchers described the thought chain as "a series of intermediate natural language reasoning steps leading to the final result."

OpenAI has explained the technique as o1 "learns to break down tricky steps into simpler ones. It learns to try another approach when the current one doesn't work. This process dramatically improves the model's ability to reason."