OpenAI updates its smarter AI model with improved reasoning skills

OpenAI today announced an improved version of its most capable artificial intelligence model to date – one that takes even more time to deliberate on the questions – only one day later Google announces its first model of this type.

OpenAI’s new model, called o3, replaces o1, which the company introduced in September. Like the o1, the new model spends time ruminating on a problem to give better answers to questions that require step-by-step logical reasoning. (OpenAI chose to skip the “o2” moniker because it’s already the name of a mobile carrier in the UK.)

“We see this as the beginning of the next phase of AI,” OpenAI CEO Sam Altman said in a livestream on Friday. “Where you can use these models to do increasingly complex tasks that require a lot of reasoning.”

The o3 model scores much higher on several measures than its predecessor, OpenAI says, including those measuring complex coding skills and advanced math and science skills. It is three times better than o1 at answering questions ARC-AGIa benchmark designed to test the ability of AI models to reason about extremely difficult mathematical and logical problems encountered for the first time.

Google is pursuing a similar line of research. Noam Shazeer, a Google researcher, yesterday revealed in a post on X that the company has developed its own reasoning model, called Gemini 2.0 Flash Thinking. Google CEO Sundar Pichai called it “our most thoughtful model ever” in his own place.

The two dueling models show the competition between OpenAI and Google to be fiercer than ever. It is crucial for OpenAI to demonstrate that it can continue to make progress as it seeks to attract more investment and build a profitable business. Google meanwhile is desperate to show it remains at the forefront of AI research.

The new models also show how AI companies are increasingly looking beyond the simple scaling of AI models to unlock greater intelligence.

OpenAI says there are two versions of the new model, o3 and o3-mini. The company is not yet making the models publicly available, but says it will invite outsiders to apply to test them. OpenAI today also revealed more details of the techniques used to align o1. This involves having the model reason on the nature of the request that is given to question whether it can contravene its guardrails.