OpenAI launches new 'reasoning' model, o3-mini

OpenAI on Friday launched a new artificial intelligence (AI) "reasoning" model, o3-mini.

The company said that the latest addition to its o family of reasoning models is both "powerful" and "affordable."

"Today's launch marks ... an important step towards broadening accessibility to advanced AI in service of our mission," an OpenAI spokesperson said.

OpenAI first previewed the model in December, alongside the more capable o3 system.

The o3-mini model is fine-tuned for STEM problems, specifically in programming, math and science. While its capabilities are largely on par with the o1 family – o1 and o1-mini – it runs faster and costs less, according to OpenAI.

The company claimed that o3-mini made 39 percent fewer "major mistakes" on "tough real-world questions" in A/B tests compared to o1-mini, with "clearer" and "faster" responses.

Starting Friday, o3-mini is available to all users via ChatGPT and to select developers via OpenAI's API, although it does not yet support image analysis.

"While o1 remains our broader general-knowledge reasoning model, o3-mini provides a specialized alternative for technical domains requiring precision and speed," OpenAI wrote in a blog post on Friday. "The release of o3-mini marks another step in OpenAI's mission to push the boundaries of cost-effective intelligence."

According to a report by TechCrunch, o3-mini is not OpenAI's most powerful model to date, nor does it outperform DeepSeek's R1 reasoning model in every benchmark.