OpenAI Revives GPT-4o After GPT-5 Backlash

OpenAI just brought GPT-4o back after an avalanche of complaints about its shiny new GPT-5. The newer model, hyped as the “smartest, fastest” yet, didn’t land well — users said it felt “sterile,” “dry,” and way less friendly.

CEO Sam Altman admitted they “underestimated” how much people valued GPT-4o’s personality, even if GPT-5 scores better in performance.

Updates to ChatGPT:

You can now choose between “Auto”, “Fast”, and “Thinking” for GPT-5. Most users will want Auto, but the additional control will be useful for some people.

Rate limits are now 3,000 messages/week with GPT-5 Thinking, and then extra capacity on GPT-5 Thinking…

— Sam Altman (@sama) August 13, 2025

Why GPT-5 Didn’t Click

GPT-5 uses a real-time router to switch between quick, efficient answers and deeper reasoning. Sounds great on paper, but for many, it lost the charm. Redditors said it was “emotionally distant” and “too businesslike,” even after customizing instructions.

“Sure, 5 is fine—if you hate nuance and feeling things,” one user quipped.

What’s Next

For now, GPT-4o stays available for ChatGPT Plus users, but there’s no promise it’ll be around forever. OpenAI says it will “watch usage” before deciding. Meanwhile, they’re working to make GPT-5 “warmer” and doubling its rate limits for Plus users.

The whole saga has reignited talk about how attached people can get to chatbots — and whether chasing “smarter” AIs means losing the human touch people secretly love.

This article was previously published on bahrainmoments. To see the original article, click here