MiniMax to Open-Source 2.7 Trillion Parameter AI Model
Curated by the Inblix editorial team
Chinese AI startup MiniMax is developing a massive new language model with 2.7 trillion parameters, internally named M3 Pro (though the name might change). The company plans to open-source the model, potentially as early as the third quarter of this year. That would make it the largest Chinese AI model ever released—far bigger than MiniMax’s current top model, which has 428 billion parameters. Bigger models generally perform better on complex reasoning tasks and multi-step instructions. Chinese open-source models have already become popular among developers, especially for cost-effective, high-volume applications. But recent reports suggest the Chinese government may tighten controls on future AI releases. Why it matters: If MiniMax follows through, it could shake up the global open-source AI landscape by offering a frontier-scale model for free, forcing competitors to rethink their strategies and potentially escalating the open-versus-closed AI debate.
MiniMax competes with Zhipu, DeepSeek, and Moonshot AI, so this move could reshape the Chinese AI market.
💡 Key Takeaways
- MiniMax plans to open-source a 2.7 trillion parameter model called M3 Pro, potentially by Q3 2025.
- The model would be significantly larger than current Chinese AI models, improving complex reasoning capabilities.
- Chinese government tightening regulations on open-source AI could impact the release or future development of such models.
Keep reading: See related articles below for more coverage on this topic.
Get smarter about AI
The sharpest AI news, curated daily. Delivered free to your inbox.