Andrew Chen Reviews GLM 5.2

andrewchen · x · 2026-07-13

Andrew Chen shared his experience playing with GLM 5.2 over the past week, calling it highly competitive. He remarked that if he could put it in a desktop box, he might not need anything else.

He later added a more cautious perspective: while every generation of frontier or adjacent LLMs makes us feel they are "already enough," humans will keep inventing new compute-dependent applications, meaning we will endlessly chase the next, more powerful model.

Related event: GLM 5.2 Impresses Testers, but Compute Demands Grow(2 posts)→

Original post →

More from Models

Models channel →