 |
Users worldwide can use both versions of ChatGPT starting today. Photo: ZDnet . |
On the morning of April 17th (Vietnam time), just two days after announcing GPT-4.1, OpenAI officially launched not one, but two new models, named o3 and o4-mini. Both models demonstrate superior reasoning capabilities with many powerful improvements.
OpenAI states that o3 is their most advanced reasoning model to date, performing strongly in tasks related to programming, mathematics, and science . o4-mini is considered a more cost-effective option that still delivers “impressive results” in similar fields.
More notably, both models can utilize and combine all the tools available in ChatGPT, including web browsers and image generation. This is a new capability not seen in previous OpenAI systems, allowing for more efficient solutions to complex, multi-step problems.
At the same time, o3 and o4-mini not only receive images, but can also interpret what they “see.” The models can understand images of whiteboards, diagrams, or sketches, even those of poor quality. They also have the ability to adjust the images as part of the reasoning process.
OpenAI also launched a new programming agent (similar to Claude Code) called Codex CLI. This tool is designed to provide programmers with a minimalist interface, allowing them to connect OpenAI models to source code on their local machine. Codex CLI is already compatible with o3 and o4-mini, and will soon support GPT-4.1 as well.
In February, OpenAI CEO Sam Altman stated that o3 would not be released as a standalone product but would be integrated into GPT-5. However, in early April, he announced a change in plans.
“There are many reasons for this,” Altman wrote on X. The integration proved more difficult than he had anticipated. However, he believes the “GPT-5 version is much better than initially expected,” and anticipates its release in the next few months.
The o3 and o4-mini versions are now available to users of ChatGPT Plus, Pro, and Team plans starting today. In the coming weeks, OpenAI will officially launch o3-pro, an even more powerful version of its core reasoning model, and offer it to users of the Pro plan. In the meantime, users can continue to use o1-pro as before.
Source: https://znews.vn/diem-hay-nhat-cua-2-mo-hinh-chatgpt-vua-ra-mat-post1546596.html
Comment (0)