Openai presented GPT-5 on Thursday, the new version of its Chatbot Chatgpt, capable of performing tasks and making decisions without human intervention, marking an advance towards autonomous artificial intelligence.
Deployed today, GPT-5 introduces “agentive” capabilities that allow you according to Euro News.
This model, available to all Chatgpt users, including free plan, improves response accuracy, reduces inventions in consultations without sufficient data and optimizes answers to health questions, identifying possible medical problems.
The company stressed that these improvements were achieved after intensive training, although it did not specify the technical details of the process.
The launch, anticipated after an accidental filtration in Github, positions GPT-5 as a key step towards the “agent” AI, capable of executing complex tasks autonomously, such as looking for information or solving problems in work environments.
Sam Altman, CEO of Openai, said in a June publication that these agents could act as “virtual co -workers” in roles such as the engineering of softwareperforming tasks that a professional with a few years of experience could handle, although still under human supervision.
However, the text does not detail the level of real autonomy or the limits of these capacities, which could generate questions about their reliability in critical scenarios.
Unlike previous models, OpenAI released the GPT-5 source code, the first from GPT-2 in 2019, allowing custom users.
This launch follows the GPT-Oss series, presented days before, and competes with Microsoft and Google AI agents.
It promises to transform sectors such as customer service, lack of information on possible ethical or security risks. It only remains to wait.
