X · @teortaxesTex
· X / Twitter
> when it comes to writing, GLM5.2's abilities fall way short of Deepseek V4 Pro—not by a little, but by a mile. One's a Tsinghua basement-dweller, t…
> when it comes to writing, GLM5.2's abilities fall way short of Deepseek V4 Pro—not by a little, but by a mile. One's a Tsinghua basement-dweller, the other's a Peking University prodigythe power of scale…Michael Anti: GLM5.2在编程上毫无疑问是能打，准AGI水平，你看看智谱套餐10点永远抢不到就知道了。但在写作方面，GLM5.2的能力比Deepseek V4 Pro差的不是一点半点，一个是清华宅男，一个是北大才