Paul Couvert(@itsPaulAi)

Ok this small local model is sooo good MiniCPM5-2B has only 2B parameters (!!) and can run agents o...

8.5内容质量

TL;DR · AI 摘要

MiniCPM5-2B模型以2B参数实现高性能，适合本地部署且开源，性能超越更大模型。

核心要点

  • MiniCPM5-2B仅2B参数，可在2GB RAM设备上运行
  • 在Artificial Analysis Intelligence Index中得分23，4B以下模型排名第一
  • Tsinghua University提出Densing Law：模型能力密度每3.5个月翻倍

结构提纲

按章节快速跳转。

  1. §模型介绍

    介绍MiniCPM5-2B的参数规模与本地运行能力

  2. ·性能验证

    展示模型在Hugging Face任务中的实际执行效果

  3. ›评估数据

    引用Artificial Analysis Intelligence Index v4.1.1排名结果

  4. ·Densing Law

    解释模型能力密度增长规律及研究背景

  5. ›开源策略

    说明OpenBMB开放的训练方法与数据资源

  6. ·技术对比

    对比MiniCPM5-2B与6倍参数模型的评估表现

思维导图

用一张图看清主题之间的关系。

查看大纲文本（无障碍 / 无 JS 友好）
  • MiniCPM5-2B模型分析
    • 模型特性
      • 2B参数/2GB RAM运行
      • 开源授权
    • 性能表现
      • Hugging Face任务执行
      • Artificial Analysis Index #1
    • 技术突破
      • Densing Law发现
      • 能力密度增长规律

金句 / Highlights

值得收藏与分享的关键句。

#MiniCPM5-2B#开源模型#Densing Law#本地部署
打开原文

Paul Couvert on X: "Ok this small local model is sooo good MiniCPM5-2B has only 2B parameters (!!) and can run agents on any laptop or even your phone fully offline. This 100% open source model can perform coding/agentic/tool use tasks and can run on just 2GB RAM! And it's genuinely good even in… / X

Paul Couvert

@itsPaulAi

Ok this small local model is sooo good MiniCPM5-2B has only 2B parameters (!!) and can run agents on any laptop or even your phone fully offline. This 100% open source model can perform coding/agentic/tool use tasks and can run on just 2GB RAM! And it's genuinely good even in Hermes agent! I asked it to: - Go to Hugging Face - Find the Models section - Evaluate models across multiple criteria - Create a CSV with the top 15 It did it! In the Artificial Analysis Intelligence Index v4.1.1 results cited in the official release, MiniCPM5-2B scores 23 and ranks #1 among open-source models under 4B parameters. So it's not about the size of the model anymore but way more the density of intelligence: Researchers from Tsinghua University and ModelBest proposed the “Densing Law”: the maximum capability density of open-source pretrained base models roughly doubled every 3.5 months over the period studied! And what's also interesting is that OpenBMB's open-source approach goes beyond just releasing model weights with - Selected training methods - Agent-related data - Data refinement resources Also being opened up, including the RL stack with Meshy + JustRL II. This model even outperforms some models around 6x larger on selected evaluations such as GDPval-AA v2!

$

00:00

/$

5:24 PM · Sep 9, 2026

14.6K

Views

15

27

232

324