Just two days after the announcement of GPT-4.1, OpenAI is releasing not one, but two new models at once. Today, the company announced the public availability of o3 and o4-mini. OpenAI claims that o3 is its most advanced reasoning model, which demonstrates “high performance” in coding, math, and science tasks. As for the o4-mini, OpenAI positions it as a cheaper alternative that still demonstrates “impressive results” in the same areas.
Moreover, both models offer new features that were not available in previous OpenAI systems. For the first time, the company’s reasoning models can utilize and combine all the tools available in ChatGPT, including web browsing and image generation. The company claims that this capability allows o3 and o4-mini to more effectively solve complex, multi-step tasks and “take real steps toward acting independently.”
At the same time, o3 and o4-mini can not only see images, but also interpret and “think” about them, which greatly expands their visual processing capabilities. For example, you can upload images of whiteboards, diagrams, or sketches-even poor quality ones-and the new models will understand them. They can also adjust images during the reasoning process.
“The combination of state-of-the-art reasoning with full access to tools delivers significantly higher performance in academic tests and real-world tasks, setting a new standard for both intelligence and utility,” OpenAI says.
Separately, OpenAI is releasing a new coding agent (a la Claude Code) called Codex CLI. It is designed to provide developers with a minimal interface that they can use to link OpenAI models to their local code. Out of the box, it works with o3 and o4-mini, with support for GPT-4.1 planned for the near future.
Today’s announcement comes after OpenAI CEO Sam Altman said the company was changing course, which he detailed in February. At the time, Altman said that OpenAI would not be releasing o3, which the company first unveiled late last year, as a standalone product. However, in early April, he announced a “change of plans,” noting that OpenAI was moving forward with the release of o3 and o4-mini.
“There are many reasons for this, but the most exciting one is that we will be able to make GPT-5 much better than we originally planned,” he wrote on X. “We’ve also found that seamless integration has proven to be more challenging than we thought, and we want to make sure we have enough capacity to support what we expect to be unprecedented demand.”
This means that the optimizations Altman promised in February will likely have to wait until at least the release of GPT-5, which he said will be coming sometime in the next “few months.”
In the meantime, ChatGPT Plus, Pro, and Team users can start using o3 and o4-mini today. In the next few weeks, OpenAI will release o3-pro, an even more powerful version of its flagship reasoning model, and make it available to Pro subscribers. In the meantime, these users can continue to use o1-pro.
