Chinese AI labs delay open-source releases for Minimax, GLM, Qwen, and Mimo models
Multiple top Chinese AI labs simultaneously delay releasing their latest open-source models, sparking speculation.
A coordinated delay in open-source releases from China's leading AI labs has sparked intense speculation within the tech community. Companies including Minimax (Minimax-m2.7), Zhipu AI (GLM-5.1/5-turbo/5v-turbo), Qwen (Qwen3.6), and Mimo (Mimo-v2-pro) have all postponed the public release of their latest models, each providing nearly identical justifications about needing more time for improvements and promising a future release. The synchronous nature of these announcements, occurring within a tight timeframe, has struck many observers as inorganic, suggesting possible external coordination or a shared strategic pivot.
This pattern raises critical questions about the future of open-source AI from China, a region that has been a significant contributor to the global ecosystem. The concern is that these delays may represent a gradual transition toward closed-source models, where initial postponements of '2-3 weeks' could evolve into indefinite holds for subsequent versions. Such a shift would mark a significant departure from recent practices and could impact global research, development, and the competitive landscape by restricting access to state-of-the-art Chinese AI advancements.
- Minimax, Zhipu AI (GLM), Qwen, and Mimo have all delayed open-sourcing their latest flagship model versions simultaneously.
- Each company is using similar language, citing model 'improvements' as the reason for the delay with vague future release promises.
- The coordinated pattern has led to widespread speculation about a strategic industry shift towards keeping advanced AI models proprietary.
Why It Matters
A move away from open-source could limit global access to cutting-edge AI, reshaping research and commercial competition.