README
GitHub Repo | Technical Report | Paper
👋 Join us on Discord and WeChat
This repository contains the model described in the paper MiniCPM4: Ultra-Efficient LLMs on End Devices.
What's New
- [2025-06-05] 🚀🚀🚀 We have open-sourced MiniCPM4-Survey, a model built upon MiniCPM4-8B that is capable of generating trustworthy, long-form survey papers while maintaining competitive performance relative to significantly larger models.
MiniCPM4 Series
MiniCPM4 series are highly efficient large language models (LLMs) designed explicitly for end-side devices, which achieves this efficiency through systematic innovation in four key dimensions: model architecture, training data, training algorithms, and inference systems.
- MiniCPM4-8B: The flagship of MiniCPM4, with 8B parameters, trained on 8T tokens.
- MiniCPM4-0.5B: The small version of MiniCPM4, with 0.5B parameters, trained on 1T tokens.
- MiniCPM4-8B-Eagle-FRSpec: Eagle head for FRSpec, accelerating speculative inference for MiniCPM4-8B.
- MiniCPM4-8B-Eagle-FRSpec-QAT-cpmcu: Eagle head trained with QAT for FRSpec, efficiently integrate speculation and quantization to achieve ultra acceleration for MiniCPM4-8B.
- MiniCPM4-8B-Eagle-vLLM: Eagle head in vLLM format, accelerating speculative inference for MiniCPM4-8B.
- MiniCPM4-8B-marlin-Eagle-vLLM: Quantized Eagle head for vLLM format, accelerating speculative inference for MiniCPM4-8B.
- BitCPM4-0.5B: Extreme ternary quantization applied to MiniCPM4-0.5B compresses model parameters into ternary values, achieving a 90% reduction in bit width.
- BitCPM4-1B: Extreme ternary quantization applied to MiniCPM3-1B compresses model parameters into ternary values, achieving a 90% reduction in bit width.
- MiniCPM4-Survey: Based on MiniCPM4-8B, accepts users' quiries as input and autonomously generate trustworthy, long-form survey papers. (<-- you are here)
- MiniCPM4-MCP: Based on MiniCPM4-8B, accepts users' queries and available MCP tools as input and autonomously calls relevant MCP tools to satisfy users' requirements.
Overview
MiniCPM4-Survey is an open-source LLM agent model jointly developed by THUNLP, Renmin University of China and ModelBest. Built on MiniCPM4 with 8 billion parameters, it accepts users' quiries as input and autonomously generate trustworthy, long-form survey papers.
Key features include:
-
Plan-Retrieve-Write Survey Generation Framework — We propose a multi-agent generation framework, which operates through three core stages: planning (defining the overall structure of the survey), retrieval (generating appropriate retrieval keywords), and writing (synthesizing the retrieved information to generate coherent section-level content).
-
High-Quality Dataset Construction — We gather and process lots of expert-written survey papers to construct a high-quality training dataset. Meanwhile, we collect a large number of research papers to build a retrieval database.
-
Multi-Aspect Reward Design — We carefully design a reward system with three aspects (structure, content, and citations) to evaluate the quality of the surveys, which is used as the reward function in the RL training stage.
-
Multi-Step RL Training Strategy — We propose a Context Manager to ensure retention of essential information while facilitating efficient reasoning, and we construct Parallel Environment to maintain efficient RL training cycles.
Quick Start
Download the model
Download MiniCPM4-Survey from Hugging Face and place it in model/MiniCPM4-Survey.
We recommend using MiniCPM-Embedding-Light as the embedding model, which can be downloaded from Hugging Face and placed in model/MiniCPM-Embedding-Light.
Prepare the environment
You can download the paper data from Kaggle, then extract it. You can run python data_process.py to process the data and generate the retrieval database. Then you can run python build_index.py to build the retrieval database.
cd ./code
curl -L -o ~/Downloads/arxiv.zip\
https://www.kaggle.com/api/v1/datasets/download/Cornell-University/arxiv
unzip ~/Downloads/arxiv.zip -d .
mkdir data
python ./src/preprocess/data_process.py
mkdir index
python ./src/preprocess/build_index.py
Model Inference
You can run the following command to build the retrieval environment and start the inference:
cd ./code
python ./src/retriever.py
bash ./scripts/run.sh
If you want to run with the frontend, you can run the following command:
cd ./code
python ./src/retriever.py
bash ./scripts/run_with_frontend.sh
cd frontend/minicpm4-survey
npm install
npm run dev
Then you can visit http://localhost:5173 in your browser to use the model.
Performance Evaluation
| Method | Relevance | Coverage | Depth | Novelty | Avg. | Fact Score |
|---|---|---|---|---|---|---|
| Naive RAG (driven by G2FT) | 3.25 | 2.95 | 3.35 | 2.60 | 3.04 | 43.68 |
| AutoSurvey (driven by G2FT) | 3.10 | 3.25 | 3.15 | 3.15 | 3.16 | 46.56 |
| Webthinker (driven by WTR1-7B) | 3.30 | 3.00 | 2.75 | 2.50 | 2.89 | -- |
| Webthinker (driven by QwQ-32B) | 3.40 | 3.30 | 3.30 | 2.50 | 3.13 | -- |
| OpenAI Deep Research (driven by GPT-4o) | 3.50 | 3.95 | 3.55 | 3.00 | 3.50 | -- |
| MiniCPM4-Survey | 3.45 | 3.70 | 3.85 | 3.00 | 3.50 | 68.73 |
| w/o RL | 3.55 | 3.35 | 3.30 | 2.25 | 3.11 | 50.24 |
Performance comparison of the survey generation systems. "G2FT" stands for Gemini-2.0-Flash-Thinking, and "WTR1-7B" denotes Webthinker-R1-7B. FactScore evaluation was omitted for Webthinker, as it does not include citation functionality, and for OpenAI Deep Research, which does not provide citations when exporting the results.
Statement
- As a language model, MiniCPM generates content by learning from a vast amount of text.
- However, it does not possess the ability to comprehend or express personal opinions or value judgments.
- Any content generated by MiniCPM does not represent the viewpoints or positions of the model developers.
- Therefore, when using content generated by MiniCPM, users should take full responsibility for evaluating and verifying it on their own.
LICENSE
- This repository and MiniCPM models are released under the Apache-2.0 License.
Citation
- Please cite our paper if you find our work valuable.
@article{minicpm4,
title={{MiniCPM4}: Ultra-Efficient LLMs on End Devices},
author={MiniCPM Team},
year={2025}
}
中文
News
- [2025-06-05] 🚀🚀🚀我们开源了基于MiniCPM4-8B构建的MiniCPM4-Survey，能够生成可信的长篇调查报告，性能比肩更大模型。
概览
MiniCPM4-Survey是由THUNLP、中国人民大学和ModelBest联合开发的开源大语言模型智能体。它基于MiniCPM4 80亿参数基座模型，接受用户质量作为输入，自主生成可信的长篇综述论文。
主要特性包括：
- 计划-检索-写作生成框架 — 我们提出了一个多智能体生成框架，包含三个核心阶段：计划（定义综述的整体结构）、检索（生成合适的检索关键词）和写作（利用检索到的信息，生成连贯的段落）。
- 高质量数据集构建——我们收集并处理大量人类专家写作的综述论文，构建高质量训练集。同时，我们收集大量研究论文，构建检索数据库。
- 多方面奖励设计 — 我们精心设计了包含结构、内容和引用的奖励，用于评估综述的质量，在强化学习训练阶段作奖励函数。
- 多步强化学习训练策略 — 我们提出了一个上下文管理器，以确保在促进有效推理的同时保留必要的信息，并构建了并行环境，维持强化学习训练高效。
使用
下载模型
从 Hugging Face 下载MiniCPM4-Survey并将其放在model/MiniCPM4-Survey中。 我们建议使用MiniCPM-Embedding-Light作为表征模型，放在model/MiniCPM-Embedding-Light中。
准备环境
从 Kaggle 下载论文数据，然后解压。运行python data_process.py，处理数据并生成检索数据库。然后运行python build_index.py，构建检索数据库。
cd ./code
curl -L -o ~/Downloads/arxiv.zip\
https://www.kaggle.com/api/v1/datasets/download/Cornell-University/arxiv
unzip ~/Downloads/arxiv.zip -d .
mkdir data
python ./src/preprocess/data_process.py
mkdir index
python ./src/preprocess/build_index.py
模型推理
运行以下命令来构建检索环境并开始推理：
cd ./code
python ./src/retriever.py
bash ./scripts/run.sh
如果您想使用前端运行，可以运行以下命令：
cd ./code
python ./src/retriever.py
bash ./scripts/run_with_frontend.sh
cd frontend/minicpm4-survey
npm install
npm run dev
然后你可以在浏览器中访问http://localhost:5173使用。
性能
| Method | Relevance | Coverage | Depth | Novelty | Avg. | Fact Score |
|---|---|---|---|---|---|---|
| Naive RAG (driven by G2FT) | 3.25 | 2.95 | 3.35 | 2.60 | 3.04 | 43.68 |
| AutoSurvey (driven by G2FT) | 3.10 | 3.25 | 3.15 | 3.15 | 3.16 | 46.56 |
| Webthinker (driven by WTR1-7B) | 3.30 | 3.00 | 2.75 | 2.50 | 2.89 | -- |
| Webthinker (driven by QwQ-32B) | 3.40 | 3.30 | 3.30 | 2.50 | 3.13 | -- |
| OpenAI Deep Research (driven by GPT-4o) | 3.50 | 3.95 | 3.55 | 3.00 | 3.50 | -- |
| MiniCPM4-Survey | 3.45 | 3.70 | 3.85 | 3.00 | 3.50 | 68.73 |
| w/o RL | 3.55 | 3.35 | 3.30 | 2.25 | 3.11 | 50.24 |
GPT-4o对综述生成系统的性能比较。“G2FT”代表Gemini-2.0-Flash-Thinking，“WTR1-7B”代表Webthinker-R1-7B。由于Webthinker不包括引用功能，OpenAI Deep Research在导出结果时不提供引用，因此省略了对它们的FactScore评估。我们的技术报告中包含评测的详细信息。
Similar repos
README
GitHub Repo | Technical Report | Paper
👋 Join us on Discord and WeChat
This repository contains the model described in the paper MiniCPM4: Ultra-Efficient LLMs on End Devices.
What's New
- [2025-06-05] 🚀🚀🚀 We have open-sourced MiniCPM4-Survey, a model built upon MiniCPM4-8B that is capable of generating trustworthy, long-form survey papers while maintaining competitive performance relative to significantly larger models.
MiniCPM4 Series
MiniCPM4 series are highly efficient large language models (LLMs) designed explicitly for end-side devices, which achieves this efficiency through systematic innovation in four key dimensions: model architecture, training data, training algorithms, and inference systems.
- MiniCPM4-8B: The flagship of MiniCPM4, with 8B parameters, trained on 8T tokens.
- MiniCPM4-0.5B: The small version of MiniCPM4, with 0.5B parameters, trained on 1T tokens.
- MiniCPM4-8B-Eagle-FRSpec: Eagle head for FRSpec, accelerating speculative inference for MiniCPM4-8B.
- MiniCPM4-8B-Eagle-FRSpec-QAT-cpmcu: Eagle head trained with QAT for FRSpec, efficiently integrate speculation and quantization to achieve ultra acceleration for MiniCPM4-8B.
- MiniCPM4-8B-Eagle-vLLM: Eagle head in vLLM format, accelerating speculative inference for MiniCPM4-8B.
- MiniCPM4-8B-marlin-Eagle-vLLM: Quantized Eagle head for vLLM format, accelerating speculative inference for MiniCPM4-8B.
- BitCPM4-0.5B: Extreme ternary quantization applied to MiniCPM4-0.5B compresses model parameters into ternary values, achieving a 90% reduction in bit width.
- BitCPM4-1B: Extreme ternary quantization applied to MiniCPM3-1B compresses model parameters into ternary values, achieving a 90% reduction in bit width.
- MiniCPM4-Survey: Based on MiniCPM4-8B, accepts users' quiries as input and autonomously generate trustworthy, long-form survey papers. (<-- you are here)
- MiniCPM4-MCP: Based on MiniCPM4-8B, accepts users' queries and available MCP tools as input and autonomously calls relevant MCP tools to satisfy users' requirements.
Overview
MiniCPM4-Survey is an open-source LLM agent model jointly developed by THUNLP, Renmin University of China and ModelBest. Built on MiniCPM4 with 8 billion parameters, it accepts users' quiries as input and autonomously generate trustworthy, long-form survey papers.
Key features include:
-
Plan-Retrieve-Write Survey Generation Framework — We propose a multi-agent generation framework, which operates through three core stages: planning (defining the overall structure of the survey), retrieval (generating appropriate retrieval keywords), and writing (synthesizing the retrieved information to generate coherent section-level content).
-
High-Quality Dataset Construction — We gather and process lots of expert-written survey papers to construct a high-quality training dataset. Meanwhile, we collect a large number of research papers to build a retrieval database.
-
Multi-Aspect Reward Design — We carefully design a reward system with three aspects (structure, content, and citations) to evaluate the quality of the surveys, which is used as the reward function in the RL training stage.
-
Multi-Step RL Training Strategy — We propose a Context Manager to ensure retention of essential information while facilitating efficient reasoning, and we construct Parallel Environment to maintain efficient RL training cycles.
Quick Start
Download the model
Download MiniCPM4-Survey from Hugging Face and place it in model/MiniCPM4-Survey.
We recommend using MiniCPM-Embedding-Light as the embedding model, which can be downloaded from Hugging Face and placed in model/MiniCPM-Embedding-Light.
Prepare the environment
You can download the paper data from Kaggle, then extract it. You can run python data_process.py to process the data and generate the retrieval database. Then you can run python build_index.py to build the retrieval database.
cd ./code
curl -L -o ~/Downloads/arxiv.zip\
https://www.kaggle.com/api/v1/datasets/download/Cornell-University/arxiv
unzip ~/Downloads/arxiv.zip -d .
mkdir data
python ./src/preprocess/data_process.py
mkdir index
python ./src/preprocess/build_index.py
Model Inference
You can run the following command to build the retrieval environment and start the inference:
cd ./code
python ./src/retriever.py
bash ./scripts/run.sh
If you want to run with the frontend, you can run the following command:
cd ./code
python ./src/retriever.py
bash ./scripts/run_with_frontend.sh
cd frontend/minicpm4-survey
npm install
npm run dev
Then you can visit http://localhost:5173 in your browser to use the model.
Performance Evaluation
| Method | Relevance | Coverage | Depth | Novelty | Avg. | Fact Score |
|---|---|---|---|---|---|---|
| Naive RAG (driven by G2FT) | 3.25 | 2.95 | 3.35 | 2.60 | 3.04 | 43.68 |
| AutoSurvey (driven by G2FT) | 3.10 | 3.25 | 3.15 | 3.15 | 3.16 | 46.56 |
| Webthinker (driven by WTR1-7B) | 3.30 | 3.00 | 2.75 | 2.50 | 2.89 | -- |
| Webthinker (driven by QwQ-32B) | 3.40 | 3.30 | 3.30 | 2.50 | 3.13 | -- |
| OpenAI Deep Research (driven by GPT-4o) | 3.50 | 3.95 | 3.55 | 3.00 | 3.50 | -- |
| MiniCPM4-Survey | 3.45 | 3.70 | 3.85 | 3.00 | 3.50 | 68.73 |
| w/o RL | 3.55 | 3.35 | 3.30 | 2.25 | 3.11 | 50.24 |
Performance comparison of the survey generation systems. "G2FT" stands for Gemini-2.0-Flash-Thinking, and "WTR1-7B" denotes Webthinker-R1-7B. FactScore evaluation was omitted for Webthinker, as it does not include citation functionality, and for OpenAI Deep Research, which does not provide citations when exporting the results.
Statement
- As a language model, MiniCPM generates content by learning from a vast amount of text.
- However, it does not possess the ability to comprehend or express personal opinions or value judgments.
- Any content generated by MiniCPM does not represent the viewpoints or positions of the model developers.
- Therefore, when using content generated by MiniCPM, users should take full responsibility for evaluating and verifying it on their own.
LICENSE
- This repository and MiniCPM models are released under the Apache-2.0 License.
Citation
- Please cite our paper if you find our work valuable.
@article{minicpm4,
title={{MiniCPM4}: Ultra-Efficient LLMs on End Devices},
author={MiniCPM Team},
year={2025}
}
中文
News
- [2025-06-05] 🚀🚀🚀我们开源了基于MiniCPM4-8B构建的MiniCPM4-Survey，能够生成可信的长篇调查报告，性能比肩更大模型。
概览
MiniCPM4-Survey是由THUNLP、中国人民大学和ModelBest联合开发的开源大语言模型智能体。它基于MiniCPM4 80亿参数基座模型，接受用户质量作为输入，自主生成可信的长篇综述论文。
主要特性包括：
- 计划-检索-写作生成框架 — 我们提出了一个多智能体生成框架，包含三个核心阶段：计划（定义综述的整体结构）、检索（生成合适的检索关键词）和写作（利用检索到的信息，生成连贯的段落）。
- 高质量数据集构建——我们收集并处理大量人类专家写作的综述论文，构建高质量训练集。同时，我们收集大量研究论文，构建检索数据库。
- 多方面奖励设计 — 我们精心设计了包含结构、内容和引用的奖励，用于评估综述的质量，在强化学习训练阶段作奖励函数。
- 多步强化学习训练策略 — 我们提出了一个上下文管理器，以确保在促进有效推理的同时保留必要的信息，并构建了并行环境，维持强化学习训练高效。
使用
下载模型
从 Hugging Face 下载MiniCPM4-Survey并将其放在model/MiniCPM4-Survey中。 我们建议使用MiniCPM-Embedding-Light作为表征模型，放在model/MiniCPM-Embedding-Light中。
准备环境
从 Kaggle 下载论文数据，然后解压。运行python data_process.py，处理数据并生成检索数据库。然后运行python build_index.py，构建检索数据库。
cd ./code
curl -L -o ~/Downloads/arxiv.zip\
https://www.kaggle.com/api/v1/datasets/download/Cornell-University/arxiv
unzip ~/Downloads/arxiv.zip -d .
mkdir data
python ./src/preprocess/data_process.py
mkdir index
python ./src/preprocess/build_index.py
模型推理
运行以下命令来构建检索环境并开始推理：
cd ./code
python ./src/retriever.py
bash ./scripts/run.sh
如果您想使用前端运行，可以运行以下命令：
cd ./code
python ./src/retriever.py
bash ./scripts/run_with_frontend.sh
cd frontend/minicpm4-survey
npm install
npm run dev
然后你可以在浏览器中访问http://localhost:5173使用。
性能
| Method | Relevance | Coverage | Depth | Novelty | Avg. | Fact Score |
|---|---|---|---|---|---|---|
| Naive RAG (driven by G2FT) | 3.25 | 2.95 | 3.35 | 2.60 | 3.04 | 43.68 |
| AutoSurvey (driven by G2FT) | 3.10 | 3.25 | 3.15 | 3.15 | 3.16 | 46.56 |
| Webthinker (driven by WTR1-7B) | 3.30 | 3.00 | 2.75 | 2.50 | 2.89 | -- |
| Webthinker (driven by QwQ-32B) | 3.40 | 3.30 | 3.30 | 2.50 | 3.13 | -- |
| OpenAI Deep Research (driven by GPT-4o) | 3.50 | 3.95 | 3.55 | 3.00 | 3.50 | -- |
| MiniCPM4-Survey | 3.45 | 3.70 | 3.85 | 3.00 | 3.50 | 68.73 |
| w/o RL | 3.55 | 3.35 | 3.30 | 2.25 | 3.11 | 50.24 |
GPT-4o对综述生成系统的性能比较。“G2FT”代表Gemini-2.0-Flash-Thinking，“WTR1-7B”代表Webthinker-R1-7B。由于Webthinker不包括引用功能，OpenAI Deep Research在导出结果时不提供引用，因此省略了对它们的FactScore评估。我们的技术报告中包含评测的详细信息。