Xiaomi

2026Xiaomipaid

MiMo-V2-TTS

MiMo-V2-TTS is Xiaomi's self-developed large-scale speech synthesis model, designed to provide AI agents with expressive, natural-sounding voices. It offers contextual emotion awareness, universal style adaptability, and real-time, seamless interaction, enabling agents to communicate with warmth and authenticity.

Contextual emotion awareness

Universal style adaptability

Real-time, seamless interaction

versionV2

regionChina

2026Xiaomiopen source

Xiaomi TabLDM

vInitial Release

tabular-data

foundation-model

Xiaomi-TabLDM is a tabular foundation model designed for classification and regression tasks using in-context learning. It eliminates the need for task-specific fine-tuning by leveraging large-scale synthetic pretraining on structural causal models to achieve high predictive accuracy and efficiency.

In-context learning for classification and regression without fine-tuning

Pretrained on large-scale synthetic data from structural causal models

Scikit-learn compatible API for seamless integration

pricingOpen Source (Apache License 2.0)

versionInitial Release

2026Xiaomipaid

MiMo-V2.5-Pro-UltraSpeed

vV2.5-Pro-UltraSpeed

AI language model

Xiaomi

MiMo-V2.5-Pro-UltraSpeed is an advanced AI language model developed by Xiaomi, capable of generating up to 1,200 tokens per second on a 1-trillion-parameter model. This performance is achieved through a deep collaboration with TileRT, focusing on extreme model-system co-design to optimize inference speed on commodity GPUs.

Generates up to 1,200 tokens per second on a 1-trillion-parameter model

Achieves high inference speed on commodity GPUs through model-system co-design

Utilizes FP4 quantization and DFlash speculative decoding for enhanced performance

versionV2.5-Pro-UltraSpeed

regionChina

2026Xiaomiopen source

MiMo Code

MiMo Code is an open-source AI coding assistant developed by Xiaomi, designed to enhance developer productivity by providing intelligent coding support directly within the terminal. It integrates Xiaomi's MiMo-V2.5 model and supports major large models such as DeepSeek and Kimi, offering a persistent memory system and advanced features for efficient coding tasks.

Persistent memory system

Integration with major large models (DeepSeek, Kimi, GLM)

Compose mode for complex programming tasks

pricingFree

versionV0.1.0

2026Xiaomiopen source

MiMo-V2.6 Pro & Flash

v2.6

Multimodal

The MiMo-V2.6 series is Xiaomi's latest frontier intelligence model lineup, designed to provide advanced multimodal capabilities and high-performance agentic workflows. The series includes specialized variants like the MiMo-V2.6-Pro and MiMo-V2.6-Flash, which have undergone extensive reinforcement learning training to optimize professional-grade tasks.

Frontier-level multimodal understanding

Advanced agentic reasoning for long-horizon tasks

High-throughput inference optimization

version2.6

regionChina

2026Xiaomiopen source

MiMo-V2-Omni

v2.0

Agentic AI

Coding

MiMo-V2-Omni is Xiaomi's latest multimodal foundation model designed to process and understand images, video, and audio simultaneously. It enables agents to perceive and act in the real world by integrating these modalities into a unified perceptual stream, facilitating real-time, complex decision-making across various applications.

Unified multimodal processing of images, video, and audio

Advanced perception capabilities surpassing leading models

Native support for structured tool calling and function execution

version2.0

regionChina