News for hackers. Stuff that matters.
Re:Qwen4-Max by cheema33 on Monday November 16, 2026 @02:25AM (#50514992)
> but it's clearly been incorporated now in the training data

You are not the first one to think of this. And Simon has thought of it as well, and commented on it several times.

Re:Qwen4-Max by KeplerBoy on Monday November 16, 2026 @02:33AM (#50515031)
Still, by now it's just more tradition than anything else. It's just kind of neat to have a consistent benchmark across many LLMs, even if the benchmark is completely gamed, but that is the nature of benchmarks.