Aug 21, 09:13 AM

OpenAI’s Altman Concedes GPT-5 Missteps, Vows Faster GPT-6

OpenAI released GPT-5 on 19 Aug. 2025, describing the large-language model as its most advanced yet and making it available to all ChatGPT users, with added features for Plus and Pro subscribers. The update promised sharper reasoning, coding and writing abilities, along with new health-and-safety safeguards. Two days later, Chief Executive Officer Sam Altman privately told reporters in San Francisco that the launch had alienated many of ChatGPT’s hundreds of millions of users. “I think we totally screwed up some things on the rollout,” he said, acknowledging performance complaints and frustration over the temporary removal of earlier models. Altman said OpenAI is accelerating work on GPT-6 and expects the gap between versions to be shorter than the two-plus years that separated GPT-4 and GPT-5. He cautioned, however, that persistent GPU shortages continue to limit deployment and disclosed plans to double the company’s computing capacity over the next five months. The rocky debut has renewed debate over governance and safety of frontier AI systems, with legal scholars and industry critics pointing to the episode as evidence that more external oversight may be needed before more powerful models reach the market.

#OpenAI #GPT #ChatGPT #Plus #Pro #Sam Altman #San Francisco #Altman

Written with ChatGPT .

Sources

Additional media

Image #1 for story openais-altman-concedes-gpt-5-missteps-vows-faster-gpt-6-55c5abba

Image #2 for story openais-altman-concedes-gpt-5-missteps-vows-faster-gpt-6-55c5abba

OpenAI’s Altman Concedes GPT-5 Missteps, Vows Faster GPT-6

Sources

Additional media

Similar Stories

Similar Stories