2 min Applications

OpenAI o3-pro and Google Gemini “Kingfall” on the way

OpenAI o3-pro and Google Gemini “Kingfall” on the way

OpenAI and Google are about to introduce new AI models that will further push the boundaries of advanced reasoning and programming.

OpenAI plans to release an update for ChatGPT. With this update, the company will activate the new o3 Pro model. This model has more computing power and, according to BleepingComputer , is better at thinking.

ChatGPT currently offers o4-mini, o4-mini-high, GPT-4.5, GPT-4.1, and GPT-4.1-mini, in addition to o3 for $20 Plus subscribers. The $200 Pro subscription includes all of these models plus the o1 Pro model, which provides extra computing power for more demanding tasks.

The addition of Pro means that the model has more computing power, making it better able to reliably perform complex tasks. Until now, the Pro subscription included the powerful but now outdated o1 Pro model.

The o3 Pro is set to replace this and combines the strong reasoning capabilities of o3 with the stability of o1 Pro. This offers more brainpower and better performance for difficult tasks such as advanced math, coding, and science. The official announcement of this upgrade may be on the agenda for next week.

Unknown Google model

At the same time, Google also appears to be coming up with a new AI breakthrough. Under the code name Gemini Kingfall, an unknown model briefly appeared in AI Studio for about 20 minutes before being quickly removed. According to BleepingComputer, Kingfall is a programming champion.

It does not appear to be an update of the existing Gemini 2.5-Pro model, but a completely new system that, according to early signs, distinguishes itself primarily in the area of programming. An example of its power: a Reddit user gave a simple three-line command to create a working Minecraft clone in a single HTML file, including terrain, mobs, and pixel graphics. Gemini Kingfall immediately delivered a largely functional result, although there were still a few minor flaws, such as a broken water texture.

Both developments point to a trend in which AI models are not only becoming faster and more accurate, but also more creative, more independent, and more suitable for technical and professional applications.