MiniMax Launches 2.7 Trillion Parameter Large Model

Artificial intelligence company MiniMax is developing a new large language model with a parameter count of 2.7 trillion, surpassing all domestic AI large models currently available on the market. The model is expected to be released in the third quarter of this year, with the internal code name M3 Pro; it is unclear whether this name will be retained upon official release. MiniMax plans to open-source this model, which is significantly larger than the current flagship model M3 (428 billion parameters). Larger-scale AI models are better at handling complex reasoning and multi-step instruction tasks.