On May 29, according to social media, Xiaomi announced an end-of-life notice for the MiMo-V2-Pro/Omni model, stating that the service for the old version models will be officially discontinued on June 30, 2026, and replaced by the MiMo V2.5 series. Specifically, the mimo-v2-pro will be migrated to mimo-v2.5-pro, and the mimo-v2-omni will be upgraded to the new model of the mimo-v2.5 series.

Xiaomi stated in the notice that the MiMo V2.5 series is now fully launched and running stably. This adjustment aims to encourage developers to migrate to the new version model with stronger inference capabilities and higher cost-effectiveness. The official also thanked developers for their long-term support of the MiMo open platform and advised relevant users to complete interface and business adaptation as soon as possible to avoid the impact caused by the subsequent service shutdown.

QQ20260529-142745.jpg

Looking at the product evolution path, the MiMo V2.5 series indicates that Xiaomi is further enhancing model inference performance and platform integration capabilities. As the competition of domestic large models gradually enters the stage of performance optimization and commercialization, rapid iteration of model versions and the retirement of old capabilities have become an industry norm. Compared to the early competition logic focused on parameter scale, current manufacturers emphasize inference efficiency, deployment costs, and developer ecosystem compatibility.

This MiMo model transition also reflects Xiaomi's continuous efforts to upgrade its AI open platform system. By unifying model versions and capability frameworks, it reduces maintenance costs and improves developer efficiency. In the context where AI applications are increasingly integrated into terminals and multi-scenario ecosystems, the stability and iteration pace of the model platform have become an important part of the competition among manufacturers.