OpenAI’s latest AI model, GPT-5, has been making waves in the tech world, and not just for its groundbreaking capabilities. During a recent Reddit AMA (Ask Me Anything), CEO Sam Altman addressed several concerns and quirks surrounding the new release.

The rollout wasn’t exactly smooth sailing. Users reported that GPT-5 seemed less intelligent compared to its predecessor, GPT-4o. Altman explained that a technical issue with the model’s real-time router was responsible for the perceived drop in performance. “GPT-5 will seem smarter starting today,” he promised, acknowledging that the autoswitcher was temporarily out of commission.

One of the most talked-about moments was the now-infamous “chart crime” - a presentation slide with a wildly inaccurate visual representation of benchmark scores. While Altman called it a “mega chart screwup” on social media, the published blog post showed corrected charts.

Listening to user feedback, Altman committed to several improvements. He’s considering allowing Plus subscribers to continue using GPT-4o and plans to double rate limits to give users more flexibility in exploring the new model.

The AMA highlighted OpenAI’s commitment to transparency and continuous improvement. Altman directly engaged with user concerns, promising to “continue to work to get things stable and will keep listening to feedback”.

For tech enthusiasts and AI observers, this rollout represents more than just a product launch. It’s a glimpse into the complex process of developing cutting-edge artificial intelligence, complete with its occasional stumbles and rapid iterations.

As AI continues to evolve at a breakneck pace, OpenAI remains at the forefront, pushing boundaries and openly addressing challenges along the way.