China's AI-native labs (GLM, Kimi, DeepSeek) are outpacing Tencent, Alibaba and Baidu

Antique_Atmosphere97 · reddit · 2026-11-21

A Reddit discussion highlights a contrast in China's AI race: Tencent just shipped hy4 and Alibaba has Qwen4-Max, while the author can't even name Baidu's flagship model. Yet the Chinese models people actually talk about — and use for coding and agents — mostly come from AI-native labs: Zhipu's GLM, Moonshot's Kimi, and DeepSeek.

The author asks why the incumbents, with more money, more GPUs and more users, still seem behind: organizational weight, short-term KPIs, or does model building simply not reward classic big-tech advantages?

Related event: Chinese AI-Native Labs Outshine Legacy Tech Giants(2 posts)→

Original post →

More from Companies & People

Companies & People channel →