A year ago, OpenAI's GPT-4 AI model's capacity to mimic human speech and composition caused massive buzz and existential panic.
Since then, the biggest AI question has been: When is GPT-5 coming out?
OpenAI CEO Sam Altman often answers this question in interviews and media events throughout the world with a cautious or evasive answer, occasionally promising great things.
Business Insider reports that OpenAI will release GPT-5, an enhanced AI language model that underpins ChatGPT,
in mid-2024, presumably in the summer.
Enterprise clients have gotten GPT-5 and ChatGPT upgrade demos, according to two anonymous business sources.
GPT-5 may be a “materially better” ChatGPT update this summer.
GPT-5 (or whatever it will be named) is planned to be a multimodal large language model (LLM) that accepts text or encoded visual input (called a "prompt") like its predecessor. GPT-5, like GPT-4, will be a next-token prediction model that outputs its best estimate of the most likely next token (a word fragment) in a sequence, allowing for sentence completion and code creation.