OpenAI continues its frantic race to improve its chatbot, ChatGPT. Sam Altman’s new version of generative artificial intelligence could see the light of day next summer. A handful of first testers external to OpenAI have already had access to this new version of the GPT model.
ChatGPT-5 training for this summer?
A year has passed since the arrival of ChatGPT-4. This new version of the AI model brought a considerable number of improvements, particularly with regard to the speed of execution of tasks, the complexity of the data processed and the precision of the content generated.
According to anonymous sources, relayed by the media Business Insider, the new version of ChatGPT would be more efficient than GPT-4. GPT-5 would have new features such as the execution of certain tasks in complete autonomy, where GPT-4 tends to request more details from the user.
ChatGPT-5 is still in the training phase, particularly in its security and the improvement of its “guardrails”, with the aim of avoiding the generation of illegal and immoral content.
To date, if OpenAI has not communicated a precise release date for the successor to GPT-4, various observers anticipate a release in mid-2024… However, the new version must still pass security exams before being ready to face the general public market.