“Q1: Are Chinese AI models capable?
A1: Yes. Chinese models are now close enough to the frontier that they can compete with U.S. models in many real-world tasks. GLM-5.2 is an open-weight model with roughly 750 billion total parameters and a 1 million token context window. It ranks second across other models in the front-end coding benchmark and the top among other open-weight models. Moonshot’s Kimi K2.7 Code model follows OpenAI and Anthropic models closely in the agent benchmarks and software engineering benchmark. The DeepSeek V4-Pro and Qwen3.7-Max models are also highly capable and just behind U.S.-based models. All these recent releases show a larger trend that Chinese AI models are only months, not years, behind U.S. frontier models. The Center for AI Standards and Innovation (CAISI) evaluated open-weight models and found that the gap between the leading U.S. models and DeepSeek V4 Pro is about eight months behind the leading U.S. models…”
“There are many reasons behind this trend…”
https://www.csis.org/analysis/what-know-about-chinese-ai-models