OpenAI may unveil the subsequent technology of its artificial intelligence (AI) mannequin, GPT-5 (Generative Pre-Trained Transformers 5), within the coming months, as per a brand new report. The AI mannequin will arrive because the successor to GPT-4, which was formally launched in March 2023 and at the moment powers its widespread chatbot, ChatGPT. It is claimed that the fifth iteration of GPT will even carry new capabilities equivalent to AI brokers that may function autonomously. Notably, OpenAI CTO Mira Murati additionally revealed lately that the corporate’s first text-to-video AI mannequin Sora may arrive later this 12 months, too.
According to a report by Business Insider, GPT-5 could possibly be launched round mid-year, which might be the months of June or July. The new AI mannequin was reportedly shared with a few of OpenAI’s enterprise purchasers who revealed that the brand new mannequin is extra succesful and comes with vital enhancements over its predecessor. Unnamed by the publication, the sources referred to as GPT-5 “really good” and “materially better”.
It is claimed that the brand new basis mannequin for ChatGPT will even arrive with new options. One such characteristic can be AI brokers that may be requested to carry out duties autonomously, as per the report. AI brokers look like much like GPTs – trainable mini chatbots launched in November 2023 that that may be customised for particular duties. However, unsupervised operability is a brand new ability which is reportedly being added. No additional particulars on how these AI brokers perform is understood.
The launch timeline for the AI mannequin will not be set in stone, nonetheless. As per the report, OpenAI continues to be coaching the mannequin, which implies there’s a chance of delays as properly. In an interview with Lex Fridman, CEO Sam Altman was requested about when GPT-5 could be unveiled and he replied, “I don’t know. That’s the honest answer”. But on the identical time, he highlighted that the distinction between GPT-4 and GPT-5 can be much like the distinction between GPT-3 and GPT-4.
Last week, Murati spoke about OpenAI’s one other new challenge, Sora, that may create one-minute-long movies from easy textual content prompts. As per the corporate CTO, Sora could be publicly out there within the coming months and be priced much like DALL-E.
