Xiaomi has officially started the transition from its MiMo-V2 series to the newer MiMo-V2.5 model family, marking an important update for developers using Xiaomi’s AI platform. The affected models include mimo-v2-pro, mimo-v2-omni, mimo-v2-flash, and mimo-v2-tts. According to Xiaomi’s announcement, these older model names will be fully retired at 00:00 Beijing time on June 30, 2026, meaning API calls using the old names will no longer work after that deadline.
What is changing with Xiaomi MiMo-V2 models?
Xiaomi is replacing multiple MiMo-V2 model names with newer MiMo-V2.5 versions. Before the final retirement date, the platform will use a transition period where requests sent to old model names are automatically routed to their newer equivalents.
This automatic replacement helps developers migrate gradually without immediately breaking existing applications. However, Xiaomi clearly states that after the final decommissioning time, old model names will become invalid and API requests using them will return error messages.
MiMo-V2 to MiMo-V2.5 replacement rules
The migration affects four model names across Xiaomi’s AI platform:
Xiaomi notes that mimo-v2-pro and mimo-v2-omni are fully compatible with their replacement models at the API parameter level. This means most developers should not need major code changes during the migration process, although business logic, pricing, and voice output should still be checked carefully.
Affected models and important dates
After June 30, developers must use the new MiMo-V2.5 model names directly. Any service still calling the old MiMo-V2 model names may experience request failures.
What developers need to know
The most important technical change is that old model names will no longer be treated as long-term aliases. Xiaomi is only offering temporary automatic routing during the transition window. Developers should update their API configuration, model selection logic, backend routing rules, and billing checks before the final deadline.
For speech generation, the migration from mimo-v2-tts to mimo-v2.5-tts also changes how voice timbres are mapped. The previous timbre system is no longer used directly. Instead, Xiaomi applies region-based mapping:
This means developers using text-to-speech output should verify whether the new voice mapping affects tone, pronunciation, voice consistency, or user experience in production services.
How this update improves Xiaomi’s AI platform
The MiMo-V2.5 transition gives Xiaomi a cleaner and more unified AI model structure. Instead of maintaining older V2 model names across different product categories, Xiaomi is moving developers toward a newer model family with updated routing, billing, and speech model behavior.
For users and developers, the biggest benefit is long-term stability. The automatic routing period reduces migration risk, while the final June 30 deadline gives developers a clear date for completing adaptation work. This is especially important for apps, tools, and services that depend on Xiaomi’s AI models for agent workflows, multimodal tasks, and speech generation.
