Decryption – This new language model is deployed this Thursday in Chatgpt. Far ahead of his rivals in terms of use, Openai must continue to innovate so as not to be caught up.
More powerful, faster and much smarter. Openai is not lacking in superlatives to describe GPT-5, its new language model (LLM) which will infuse all of its products this Thursday, including Chatgpt in free version. Impossible to go back after having tasted it, proclaims Sam Altman. The co -founder and director of Openai says that “GPT-5 gives for the first time the impression of talking to experts with a doctorate, who will be available at any time in your pocket”while GPT-3 (2020) was more similar to a high school student, and GPT-4 (2023) to a student.
Maths, health issues or even creative writing, this model would be better in everything and would give more the impression of interacting with a human. Users will also be able to choose the “personality” of Chatgpt between the tones “expert”, “listening”, “robot” or “cynical” – difficult in the latter case not to think of Grok, the unbridled AI of Elon …