[MiniMax May Open Source 2.7 Trillion Parameter Model in Q3, Poised to Become the World's Largest Open Source Model] According to monitoring by Beating and a report from The Information citing insiders, MiniMax is developing a 2.7 trillion parameter large language model, which could be released as early as the third quarter and is planned to be open-sourced. The model is internally codenamed M3 Pro, though the final release name has yet to be determined. If the weights are opened as planned, this model is expected to become the world's largest open-source model in terms of parameter scale. MiniMax's current flagship model, M3, has 428 billion parameters, and the new model's scale is approximately 6.3 times larger, surpassing several existing trillion-parameter open-weight models. Larger parameter scales typically benefit complex reasoning and multi-step instruction tasks, but the final performance will still depend on training quality, activation parameters, inference costs, and evaluation results. [Original Link]