Sam Altman Reveals Issues Behind GPT-5 Launch

OpenAI CEO Sam Altman has admitted that the rollout of the company’s GPT-5 model was mishandled. The AI startup faced backlash after discontinuing all its previous models immediately following the launch, prompting user complaints and subscription cancellation threats.

Key Highlights

• Sam Altman admitted, “We totally screwed up some things” during the GPT-5 rollout, acknowledging oversight.
• OpenAI quickly reinstated GPT-4o for paid users after backlash and promised warmer, smarter GPT-5 updates.

In response, OpenAI reinstated GPT-4o for ChatGPT Plus users and increased usage limits for GPT-5’s Standard and Thinking models. Although GPT-5 was promoted as a major upgrade with improved coding, reasoning, accuracy, health, and multimodal capabilities, many users reported shorter responses and less emotional depth compared to earlier versions.

Altman acknowledged the missteps during a discussion with reporters, stating, “We totally screwed up some things on the rollout.” He noted that the company underestimated how strongly users were attached to existing models and learned the difficulty of upgrading a product used by hundreds of millions overnight.

Also Read: Sam Altman Vows ‘Warmer’ GPT-5 After Complaints ChatGPT Lost Charm

Despite the criticism, OpenAI saw ChatGPT API traffic double within 48 hours of the launch, and app usage hit record levels. However, Altman admitted it was wrong to retire all older models abruptly.

OpenAI recently announced reaching 700 million weekly users — a milestone it hoped GPT-5 would accelerate further. Unlike GPT-4o’s debut earlier this year, which introduced image generation features and drove a global surge in usage, GPT-5’s launch faced mixed reception. Following user backlash, OpenAI even reinstated the model picker — a feature Sam Altman had previously vowed to remove.