On July 6th, the field of artificial intelligence witnessed a significant advancement. Meituan officially announced the open-source release of its foundational large model, LongCat-2.0, an action regarded by the industry as a major breakthrough in the application of domestic computing power in AI large models.

As the first foundational large model in the industry to be trained and inferred using domestic computing chips, LongCat-2.0 demonstrates formidable technical capabilities. The model was officially released on June 30th, and both its training and inference processes were successfully completed on a domestic computing cluster consisting of 50,000 cards, fully demonstrating the stability and high performance of domestic chips in handling trillion-parameter-level models.

To promote the inclusiveness and progress of the technology ecosystem, Meituan adopted a comprehensive open-source strategy. On July 6th, the model weights, inference engine, and detailed technical documentation of LongCat-2.0 were simultaneously launched on mainstream developer communities such as GitHub, Hugging Face, and ModelScope. Notably, the model is released under the MIT open-source license, which means developers and companies can use it for commercial purposes free of charge, significantly lowering the barrier for small and medium-sized enterprises to adopt cutting-edge AI technologies.

The domestic chip ecosystem quickly responded. On the same day Meituan announced the open source, three domestic chip manufacturers—Huawei Ascend, MoLeLineage, and Muxi Technologies—jointly announced that they had completed the inference adaptation work for the LongCat-2.0 model. This collaborative effect not only verified the compatibility of the model on domestic hardware but also marked the rapid maturity of the soft-hardware synergy ecosystem combining "domestic chips + domestic large models."

This open-source initiative is not only a concentrated demonstration of technical capabilities but also provides a solid technical foundation for the self-reliance and control of China's artificial intelligence industry through the construction of a complete open-source ecosystem. As more developers participate in the optimization and application of LongCat-2.0, the voice of China's domestic AI power in international competition is expected to further increase.