[philjohn]

> but GLM-5.2 coming 11 weeks later, scores 48%, about on par with GPT 5.5. What's next ?!

It depends how good the US labs get at stopping distillation of their models.

[HarHarVeryFunny]

So the Chinese can land a rover on Mars, are beating us back to the Moon, but can't figure out how to generate some reasoning data to train on - that's your theory ?!

[sowbug]

I suspect there's just a bit of training data on the other side of the Great Firewall that western companies can't access.