OpenAI is preparing to release new generative AI models, including an updated version of GPT-4.1, which is an improvement over the GPT-4o model. The new model is expected to be introduced alongside smaller versions—GPT-4.1 mini and nano—as early as next week. The GPT-4o model, first introduced last year, was notable for its ability to process audio, visual data, and text in real time.
Additionally, OpenAI is preparing to release the full version of the o3 reasoning model and its smaller o4 mini version, which may appear even sooner. According to information found by engineer Tibor Blaho in the code of the new version of the ChatGPT web app, the o4 mini, o4 mini high, and o3 models will soon be available to users. These models will likely become part of the ChatGPT interface and be available for selection depending on users’ subscriptions and needs.
The o3 series models focus on high performance in STEM fields (science, mathematics, programming) and logical reasoning, offering users cost-effective alternatives to larger models. The names o4-mini and o4-mini-high indicate the next generation of smaller models, which may feature adjustable levels of “reasoning effort,” allowing users to balance between performance and speed.
Although the official release date has not yet been confirmed, OpenAI CEO Sam Altman previously stated that the new o3 and o4 models are planned for release within a few weeks. These innovations may be a response to competition, particularly from Google, which recently showcased new achievements with Gemini 2.5 Pro during the Google Cloud Next event.
