OpenAI is expected to release a new AI model series, GPT-4.1, and an inference AI model named ‘o4 Mini’ next week. The GPT-4.1 series is anticipated to replace the current main model, GPT-4o, which was unveiled in May of the previous year. The delay in developing GPT-5 due to the high demand for Ghibli-style image generation has led OpenAI to introduce an interim model to counter the release of new models by competitors.
On the 13th (local time), Sam Altman, CEO of OpenAI, announced on social media platform X (formerly Twitter) that many good news updates were coming this week, starting tomorrow. Although Altman did not provide specific details, the tech industry speculates that GPT-4.1 might be unveiled as early as the 14th.
According to reports from tech media outlet The Verge, which cited engineer Thibault Vlahos, OpenAI is preparing the GPT-4.1 series. This series is expected to be the next basic model of ChatGPT, replacing the current GPT-4o and focusing on being more lightweight with different versions such as ‘Mini’ and ‘Nano’. The previously released GPT-4.5 in February demonstrated excellent performance, but its high operating costs made OpenAI prepare relatively lighter AI alternatives.
In addition to GPT-4.1, OpenAI is also working on models like o3 and o4 Mini. Initially, only the lightweight and cost-effective ‘Mini’ version of the o3 series was made publicly available. Although Altman originally planned to integrate more advanced versions of the o3 line into the future GPT-5, he recently reversed this decision, announcing an upcoming release of the o3 Pro.
OpenAI’s latest actions reflect concerns about limited computing power amidst the emergence of new competitor AI models like Google Gemini 2.5 and Meta Lama4. The company has already allocated resources intended for GPT-5 development due to the overwhelming demand for ‘ChatGPT image generation’. As a result, the release of GPT-5 has been delayed, prompting OpenAI to secure more space through lighter, latest models in the meantime.
Recent job postings hint at overcoming computing limitations as a pressing challenge for OpenAI. Altman expressed a strong desire for collaboration with individuals interested in infrastructure and large-scale computing systems, indicating an urgent need for support. He emphasized the importance of ‘optimization’, highlighting areas like compiler design and programming language development. Compilers are programs that translate code into machine language.