Updated 11 August 2025 at 11:56 IST
‘GPT-5 Seemed Way Dumber’: Sam Altman Admits New AI Model Has Faults, Promises Smarter AI and New Features
OpenAI has also improved how the system chooses which model should respond to questions, aiming to ensure users receive the most accurate answers more often.
OpenAI CEO Sam Altman acknowledged that GPT-5, the company’s latest AI model, briefly seemed “way dumber” because of a technical issue during its launch.
In an update to the Reddit community, Altman said the trouble started when the “autoswitcher,” a system that picks the best AI model to answer user questions, went offline for hours. Without it, GPT-5’s responses were noticeably less smart. “Yesterday, the autoswitcher broke and was out of commission for a chunk of the day, and the result was GPT-5 seemed way dumber,” Altman noted, adding that the glitch has been fixed.
OpenAI has also improved how the system chooses which model should respond to questions, aiming to ensure users receive the most accurate answers more often. The company plans to make this process clearer so people can see which model is replying.
The GPT-5 launch is taking longer than expected because of its large scale. API traffic has almost doubled in just 24 hours, which has slowed the full release to all users.
For ChatGPT Plus subscribers, several updates are coming:
• A redesigned interface to make it simpler to activate “thinking” mode manually.
• Double the rate limits once the rollout is complete.
• The option to keep using GPT-4o alongside GPT-5, based on user feedback.
The GPT-5 launch is one of the most significant AI upgrades this year. However, many ChatGPT fans and users seem unhappy with the new model, calling it a “disaster” and the “biggest piece of garbage even as a paid user.”
A redditor bashed GPT-5 for lacking an “essence and soul” that makes ChatGPT different from other AI bots. “GPT-5 lacks the essence and soul that separated Chatgpt (sic) from other AI bots. I sincerely wish they bring back 4o as a legacy model or something like that,” they wrote. Several ChatGPT users were also annoyed with the sudden vanishing of all other previous AI models and them forced to use the GPT-5. "I woke up this morning to find that OpenAI deleted 8 models overnight. No warning. No choice. No "legacy option," wrote a ChatGPT Plus subscriber on Reddit.
“We will continue to work to get things stable and will keep listening to feedback,” Altman promised.
Published By : Priya Pathak
Published On: 11 August 2025 at 11:56 IST