Ethan Mollick: GPT-6.2 Is OpenAI's First Frontier Release Since GPT-4.5 That Isn't a Clear Step Forward

emollick · x · 2026-11-18

Wharton professor Ethan Mollick delivered a rare negative take on OpenAI's new release: it's the first frontier launch since GPT-4.5 that isn't a clear step forward.

He's willing to bet 6.2 is amazing at some types of complex work, but in his experience and that of people he knows, it was often frustrating to use—hard to work with, needing more guidance and pushing to get things done.

In a prior post he demoed the model's impressive browser-based 'play the world' capability, even simulating zero-g physics, yet says he doesn't find himself reaching for the 6.2 models.

Original post →

More from Models

Models channel →