OpenAI is once again working on a model, reportedly called GPT-4.1. It is a successor to GPT-4o. The model will probably appear before GPT-5.0.
The Verge recently reported that OpenAI is planning to launch GPT-4.1. This is an upgrade of the existing GPT-4o model. The medium stated, based on sources familiar with the company’s plans, that OpenAI is almost ready to unveil a number of new AI models. According to informants, the company will release one of these new AI models under GPT-4.1. According to the aforementioned sources, this is a reworked version of OpenAI’s multimodal GPT-4o model.
OpenAI is testing multiple model variants
According to BleepingComputer, there are now more indications that this refreshed model is indeed on its way. As noted by AI researcher Tibor Blaho, OpenAI is already testing model variants for o3, o4-mini and GPT-4.1 (including nano and mini variants) on the OpenAI API platform. This makes it clear that GPT-4.1 exists, but it does not appear to be a successor to GPT-4.5.
According to the AI researcher, this would mean that GPT-4.1 is a successor to GPT-4o and, therefore, to the multimodal variant. GPT-4.5, on the other hand, is believed to focus more on creativity and provide better answers.
Training GPT-4 from scratch
In an OpenAI video titled Pre-Training GPT-4.5, founder and CEO Sam Altman hinted that OpenAI has a team that wants to train GPT-4 from scratch with new training data and systems. In the video, he asked the question: if everyone within the OpenAI organization could choose each employee, what would be the smallest team within OpenAI that could train GPT-4 from scratch, using everything the company knows and has? And in such a way that all systems work?
Regardless of this, it has been clear for some time that GPT-5 will not be coming onto the market for the time being. OpenAI now focuses on o3, o4-mini, o4-mini-high and GPT-4.1 (including nano and mini variants).