---
title: "Xiaomi surged about 3% during the session, announcing the open-source of its first inference large model \"Xiaomi MiMo\""
type: "News"
locale: "en"
url: "https://longbridge.com/en/news/238185824.md"
description: "Xiaomi Group surged during the trading session, currently up about 3%. On the news front, Xiaomi announced today the open-sourcing of its first large model designed for inference, \"Xiaomi MiMo,\" which integrates pre-training and post-training to comprehensively enhance inference capabilities"
datetime: "2025-04-30T02:55:48.000Z"
locales:
  - [zh-CN](https://longbridge.com/zh-CN/news/238185824.md)
  - [en](https://longbridge.com/en/news/238185824.md)
  - [zh-HK](https://longbridge.com/zh-HK/news/238185824.md)
---

> Supported Languages: [简体中文](https://longbridge.com/zh-CN/news/238185824.md) | [繁體中文](https://longbridge.com/zh-HK/news/238185824.md)


# Xiaomi surged about 3% during the session, announcing the open-source of its first inference large model "Xiaomi MiMo"

Today, Xiaomi open-sourced its first large model designed for reasoning, "Xiaomi MiMo," linking pre-training to post-training to comprehensively enhance reasoning capabilities.

In the public evaluation sets for mathematical reasoning (AIME 24-25) and coding competitions (LiveCodeBench v5), MiMo, with only 7B parameters, surpassed OpenAI's closed-source reasoning model o1-mini and Alibaba's larger open-source reasoning model QwQ-32B-Preview.

The enhancement of MiMo's reasoning capabilities is driven by innovations in data and algorithms across multiple dimensions during the pre-training and post-training phases, including:

**Pre-training: The core is to expose the model to more reasoning patterns**

-   Data: Focused on mining rich reasoning corpora and synthesizing approximately 200B tokens of reasoning data.
-   Training: Conducted three-stage training, gradually increasing training difficulty, totaling 25T tokens of training.

**Post-training: The core is an efficient and stable reinforcement learning algorithm and framework**

-   Algorithm: Proposed Test Difficulty Driven Reward to alleviate the reward sparsity problem in difficult algorithm issues and introduced Easy Data Re-Sampling strategy to stabilize RL training.
-   Framework: Designed the Seamless Rollout system, accelerating RL training by 2.29 times and validation by 1.96 times.

MiMo-7B has open-sourced 4 models to HuggingFace: https://huggingface.co/XiaomiMiMo

Technical details can be found in the technical report: \[https://github.com/XiaomiMiMo/MiMo/blob/main/MiMo-7B-Technical-Report.pdf\](https://www.oschina.net/action/GoToLink?url=https%3A%2F%2Fgithub.com%2FXiaomiMiMo%2FMiMo%2Fblob%2Fmain%2FMiMo-7B-Technical-Report.pdf)

### Related Stocks

- [XIAOMI-W (01810.HK)](https://longbridge.com/en/quote/01810.HK.md)

## Related News & Research

- [Tesla China Demand Faces More Pressure: New Rival Premium EV Hits 15,000 Orders In 34 Minutes](https://longbridge.com/en/news/280672221.md)
- [BUZZ-China's Xiaomi loses most in six months; plans $8.7 billion in AI investment](https://longbridge.com/en/news/279879816.md)
- [BREAKINGVIEWS-Apollo reaps rich rent on its Intel lifeline](https://longbridge.com/en/news/281406220.md)
- [Nvidia Stock (NVDA) Braces for a Quick Snapback after a Rare Two-Quarter Losing Streak](https://longbridge.com/en/news/281360100.md)
- [Nvidia Makes Equity Investment In Marvell Technology, Deepens NVLink Fusion Partnership](https://longbridge.com/en/news/281388869.md)