AI company MiniMax is set to release a groundbreaking large language model with 2.7 trillion parameters, according to sources. This new model, internally codenamed M3 Pro, will surpass all existing domestic AI models in size and capability. The release is anticipated as early as the third quarter of this year, with plans to open-source the model. The M3 Pro significantly exceeds the scale of MiniMax's current flagship model, M3, which has 428 billion parameters, enhancing its ability to perform complex reasoning and multi-step tasks.