Currently, seven different neural network models are available to ChatGPT users, each with its own strengths and suitable for specific scenarios. Right now, however, people have to manually select the right model, which often creates confusion, especially for those who don’t fully understand the difference between them. But it looks like that may soon change, as hinted by an OpenAI employee.
The information was revealed by a user of social network X under the nickname Roon-according to others, behind this profile is Tarun Gogineni, who is part of the technical team of OpenAI. In his dialogue with another user, he actually confirmed that the company is already working on a mechanism that will automatically match the best model to a user’s request or task.
During the discussion, one of the interlocutors mentioned that he would like to retain the ability to switch models independently. Roon replied that this option will indeed remain, so users will be able to continue to decide for themselves which neural network to use at a particular moment.
Earlier there were rumours that a similar principle of automatic selection is planned to be implemented in GPT-5. According to Yucheng Jin, co-founder and CTO of Hyperbolic Labs, the new generation of AI will be able to independently determine which of the available models will best cope with a particular task. This approach should make working with ChatGPT even easier and more convenient – especially for those who do not want to understand the technical details.