China's Open-Model Wave Broadens: GLM-5.1, Kimi K2.6, MiMo-V2.5, ERNIE-Image, Intern-S2 Weights Released
China's open-source AI model ecosystem widened its lead this month as several major labs published new weights. Zhipu open-sourced GLM-5.1, a flagship agent-engineering model strong at long-horizon coding; Moonshot released Kimi K2.6, a natively multimodal agent model; Xiaomi open-sourced MiMo-V2.5, a native full-modal model with a 1-million-token context and an agent focus; Baidu open-sourced ERNIE-Image, an 8-billion-parameter text-to-image model skilled at text rendering; and Shanghai AI Laboratory released Intern-S2-Preview, a 35-billion-parameter scientific multimodal model. The releases follow a year in which Chinese open models captured leading positions on global call rankings and Hugging Face downloads - ModelScope alone now hosts over 120,000 models and Chinese-developed models account for about 41% of platform downloads - underscoring the rising global footprint of China's open-source community.
Source: NavXD Model Index / AI Tools Intelligence, Nov 9, 2026 | 2026-11-09
- China's Open-Model Wave Broadens: GLM-5.1, Kimi K2.6, MiMo-V2.5, ERNIE-Image, Intern-S2 Weights Released
- AI Investment Narrative Shifts From Infrastructure to Intelligence Scale as Token Demand Rises
- China's AI Event Calendar Heats Up: Baidu World 2026, Nvidia AI Day Seoul, Web Summit, IC China
- Unitree Teams With Guanglun; Apple Launches HALO as Humanoid-Robot Mass Production Hits Inflection
- Shangwei's Qiyuan Robots Pass 10,000 Units; Unitree Plans JV for Oil-and-Gas Scene
- Yiwu: World's Largest Small-Commodity Hub Links 120+ Countries, Hosts 38,000 Resident Foreign Merchants
- Qwen Office 2.0 Launches as First Team-Agent Tool With Multi-User and Agent Collaboration
- China Unveils 'Nebula Constellation' Plan, Accelerating Deployment of Its Space-Computing Cloud
- Apple Intelligence Generative Service Adds iPad and Mac Scenarios, Completes Filing in China
- Web Summit 2026 Opens in Lisbon, Chinese Tech Presence Largest Ever











