Recent model releases have tightened competition among Chinese AI labs, with Xiaomi’s September launch of the open-weights MiMo-V2.6-Pro topping independent benchmarks and driving its 22.5% odds. Alibaba’s Qwen3.8 Max remains close behind on capability and leads in ecosystem scale plus download volume, supporting its 34.5% market-implied lead. Moonshot’s Kimi K3 and Z.ai’s GLM-5.3 follow at 19% and 13.5%, respectively, buoyed by strong agentic and multimodal performance plus large funding rounds. DeepSeek’s rapid iteration and low pricing keep it relevant at 8.3%, though it trails current leaders. Traders weigh open licensing, benchmark gains, and commercial traction as key swing factors heading into year-end resolution.