OpenAI Gears Up for Mid-Year Launch of GPT-5: Report

OpenAI has apparently been demonstrating GPT-5, the next generation of its large language model (LLM), to prospective buyers — and they’re very impressed with the merchandise.

One CEO told Business Insider that the LLM is notably superior, stating, “It’s really good, like materially better.” Additionally, the CEO mentioned that during the demonstration he experienced, OpenAI customized use cases and data modeling specifically for his company — and also hinted at undisclosed capabilities.

BI reports that OpenAI is considering a summer release — however, its insiders indicate that it is still undergoing training and requires “red-teaming,” a term in the tech industry referring to hiring hackers to test for vulnerabilities in one’s products.

This final aspect holds significance because, as reported by the same website nearly a year ago, GPT-4 encountered a significant issue regarding racial biases before undergoing examination by OpenAI’s skilled exploiters during the “red-teaming” process.

Although OpenAI appears to have a relaxed approach towards the LLM’s unplanned release date, complaints regarding the apparent deterioration of GPT-4 have been accumulating over the past few months as the model reaches its first anniversary. According to sources from within the company as reported by BI, there is optimism that the launch of GPT-5, along with its remarkable capabilities, will alleviate these concerns.

Certainly, even the CEO of OpenAI, Sam Altman, has engaged in criticism of his company’s most recent publicly accessible model during an extensive and comprehensive interview with MIT researcher turned podcaster, Lex Fridman.

“I think it kinda sucks,” Sam Altman told Fridman of GPT-4.

During the same interview, the podcaster inquired Altman about the possibility of providing an approximate release date, prompting the CEO to respond with somewhat cryptic answers.

“We will release an amazing new model this year,” Sam Altman said. “I don’t know what we’ll call it.”

When Fridman pressed for more details, Altman elaborated, albeit vaguely.

“We’ll be releasing various things in the upcoming months,” he elaborated. “Before we delve into discussions about a model similar to GPT-5, whether named as such or slightly inferior or superior to what one might anticipate from a GPT-5, I believe there are numerous other significant releases to prioritize.”

During the interview, Altman notably declined to respond to any inquiries regarding OpenAI’s secretive Q* project, reportedly associated with the attempted coup last November that led to his dismissal and subsequent rehiring. This adds to the mystery surrounding GPT-5, or whatever its eventual name will be.