OpenAI is developing a new artificial ielligence model called Orion is known Uil now, there have not been many explanations about the performance of this model, but now in a new report it is said that it is the next flagship artificial ielligence model OpenAI It may not jump as much as the previous models of this startup. According to this issue, it has been said that the startup maker of ChatGPT is developing new strategies in order to deal with the slowing down of the progress of its artificial ielligence models.
In a new report, OpenAI employees who tested the new model, codenamed Orion, say that while it outperforms existing models, it is less advanced than the jump from GPT-3 to GPT-4.
OpenAI’s new strategy for its AI models
In other words, the speed of improveme of OpenAI artificial ielligence models seems to be slowing down. In fact, the Orion may not be any better than its predecessors in some areas, such as coding.
To address this problem, OpenAI is said to have created a new core team whose goal is to understand how the startup can improve its models in the face of a dwindling supply of new training data. These new strategies reportedly include training Orion with artificial data, generating artificial ielligence models, and improving the models after training.
Earlier this moh, a report said that OpenAI apparely plans to release Orion by the end of December 2024, but OpenAI CEO Sam Altman denied the news.
In another report from a few mohs ago, Information Media said that this artificial ielligence surpasses other big language models in solving mathematical and programming problems.
