The OpenAI company presented a test version of the new artificial intelligence model GPT-4.5, which, according to the developers, demonstrates improved writing and programming skills and makes fewer mistakes. About this informs OpenAI press office.
The company noted that the updated model has a higher level of “emotional intelligence”, which allows it to better recognize subtle nuances in communication with users and respond to them more accurately.
An earlier version of the model, known internally by the codename Orion, is currently being tested by a limited number of software developers, as well as users of the $200-a-month ChatGPT Pro subscription.
GPT-4.5 has become the most powerful and functional language model of OpenAI. It continues the development of the line started by the GPT-4o, but focuses on versatility, going beyond highly specialized tasks in the fields of science, technology, engineering and mathematics.
Testing has shown that GPT-4.5 provides a more natural interaction with users, less often generates fabricated facts, better copes with creative tasks and demonstrates a higher ability to adapt to a variety of requests — from providing advice and psychological support to helping with household issues.
