OpenAI is gearing up to release several new AI models soon, according to sources familiar with the matter.
One of the most anticipated releases is GPT-4.1, which is expected to be an improved version of the company’s existing GPT-4o multimodal model—a system known for understanding and responding to text, images, and audio in real time.
Alongside GPT-4.1, OpenAI also plans to launch lighter versions of the model, possibly called GPT-4.1 mini and nano, as early as next week. These scaled-down models are likely designed for faster performance and lower resource use, making them suitable for smaller devices or lighter tasks.
In addition to the GPT-4.1 lineup, OpenAI is preparing to unveil its o3 reasoning model and an o4 mini variant. Recent findings in ChatGPT’s web code by AI engineer Tibor Blaho revealed mentions of o4 mini, o4 mini high, and o3, suggesting these models could go live very soon. If everything goes according to plan, both o3 and o4 mini could launch next week.
OpenAI CEO Sam Altman recently teased on X (formerly Twitter) that an exciting feature was dropping soon, though it’s unclear if that’s related to these new models. That said, OpenAI has faced capacity issues lately, which has caused delays in rolling out certain features.
Altman even warned users earlier this month to expect delays, slow service, and occasional bugs due to high demand. Just last month, the company had to limit image generation requests because of GPU overload, with Altman joking that “our GPUs are melting” from the popularity.
As of now, OpenAI hasn’t officially commented on the upcoming releases.
