README.md

---
language:
- zh
- en
library_name: transformers
license: apache-2.0
pipeline_tag: text-generation
---

<div align="center">
<img src="https://github.com/OpenBMB/MiniCPM/blob/main/assets/minicpm_logo.png?raw=true" width="500em" ></img> 
</div>

<p align="center">
<a href="https://github.com/OpenBMB/MiniCPM/" target="_blank">GitHub Repo</a> |
<a href="https://github.com/OpenBMB/MiniCPM/tree/main/report/MiniCPM_4_Technical_Report.pdf" target="_blank">Technical Report</a> 
</p>
<p align="center">
👋 Join us on <a href="https://discord.gg/3cGQn9b3YM" target="_blank">Discord</a> and <a href="https://github.com/OpenBMB/MiniCPM/blob/main/assets/wechat.jpg" target="_blank">WeChat</a>
</p>

This model is the Int4 version of MiniCPM4-0.5B, which is trained by QAT and stored in fake quantization style. It is described in the paper [MiniCPM4: Ultra-Efficient LLMs on End Devices](https://huggingface.co/papers/2506.07900).

**Abstract:**

This paper introduces MiniCPM4, a highly efficient large language model (LLM) designed explicitly for end-side devices. We achieve this efficiency through systematic innovation in four key dimensions: model architecture, training data, training algorithms, and inference systems. Specifically, in terms of model architecture, we propose InfLLM v2, a trainable sparse attention mechanism that accelerates both prefilling and decoding phases for long-context processing. Regarding training data, we propose UltraClean, an efficient and accurate pre-training data filtering and generation strategy, and UltraChat v2, a comprehensive supervised fine-tuning dataset. These datasets enable satisfactory model performance to be achieved using just 8 trillion training tokens. Regarding training algorithms, we propose ModelTunnel v2 for efficient pre-training strategy search, and improve existing post-training methods by introducing chunk-wise rollout for load-balanced reinforcement learning and data-efficient tenary LLM, BitCPM. Regarding inference systems, we propose this http URL that integrates sparse attention, model quantization, and speculative sampling to achieve efficient prefilling and decoding. To meet diverse on-device requirements, MiniCPM4 is available in two versions, with 0.5B and 8B parameters, respectively. Sufficient evaluation results show that MiniCPM4 outperforms open-source models of similar size across multiple benchmarks, highlighting both its efficiency and effectiveness. Notably, MiniCPM4-8B demonstrates significant speed improvements over Qwen3-8B when processing long sequences. Through further adaptation, MiniCPM4 successfully powers diverse applications, including trustworthy survey generation and tool use with model context protocol, clearly showcasing its broad usability.

## What's New
- [2025.06.06] **MiniCPM4** series are released! This model achieves ultimate efficiency improvements while maintaining optimal performance at the same scale! It can achieve over 5x generation acceleration on typical end-side chips! You can find technical report [here](https://github.com/OpenBMB/MiniCPM/tree/main/report/MiniCPM_4_Technical_Report.pdf).🔥🔥🔥

## 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](https://huggingface.co/openbmb/MiniCPM4-8B): The flagship of MiniCPM4, with 8B parameters, trained on 8T tokens.
- [MiniCPM4-0.5B](https://huggingface.co/openbmb/MiniCPM4-0.5B): The small version of MiniCPM4, with 0.5B parameters, trained on 1T tokens.
- [MiniCPM4-8B-Eagle-FRSpec](https://huggingface.co/openbmb/MiniCPM4-8B-Eagle-FRSpec): Eagle head for FRSpec, accelerating speculative inference for MiniCPM4-8B.
- [MiniCPM4-8B-Eagle-FRSpec-QAT-cpmcu](https://huggingface.co/openbmb/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](https://huggingface.co/openbmb/MiniCPM4-8B-Eagle-vLLM): Eagle head in vLLM format, accelerating speculative inference for MiniCPM4-8B.
- [MiniCPM4-8B-marlin-Eagle-vLLM](https://huggingface.co/openbmb/MiniCPM4-8B-marlin-Eagle-vLLM): Quantized Eagle head for vLLM format, accelerating speculative inference for MiniCPM4-8B.
- [BitCPM4-0.5B](https://huggingface.co/openbmb/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](https://huggingface.co/openbmb/BitCPM4-1B): Extreme ternary quantization applied to MiniCPM3-1B compresses model parameters into ternary values, achieving a 90% reduction in bit width.
- [MiniCPM4-Survey](https://huggingface.co/openbmb/MiniCPM4-Survey): Based on MiniCPM4-8B, accepts users' quiries as input and autonomously generate trustworthy, long-form survey papers.
- [MiniCPM4-MCP](https://huggingface.co/openbmb/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.
- [MiniCPM4-0.5B-QAT-Int4-unquantized](https://huggingface.co/openbmb/MiniCPM4-0.5B-QAT-Int4-unquantized): Int4 version of MiniCPM4-0.5B, trained by QAT and stored in fake quantization style. (**<-- you are here**)
- [MiniCPM4-0.5B-QAT-Int4-GPTQ-format](https://huggingface.co/openbmb/MiniCPM4-0.5B-QAT-Int4-GPTQ-format): Int4 version of MiniCPM4-0.5B, trained by QAT and stored in GPTQ format.
- [MiniCPM4-0.5B-QAT-Int4-GGUF](https://huggingface.co/openbmb/MiniCPM4-0.5B-QAT-Int4-GGUF): Int4 version of MiniCPM4-0.5B in GGUF.

## Introduction
MiniCPM 4 is an extremely efficient edge-side large model that has undergone efficient optimization across four dimensions: model architecture, learning algorithms, training data, and inference systems, achieving ultimate efficiency improvements.

- 🏗️ **Efficient Model Architecture:**
  - InfLLM v2 -- Trainable Sparse Attention Mechanism: Adopts a trainable sparse attention mechanism architecture where each token only needs to compute relevance with less than 5% of tokens in 128K long text processing, significantly reducing computational overhead for long texts

- 🧠 **Efficient Learning Algorithms:**
  - Model Wind Tunnel 2.0 -- Efficient Predictable Scaling: Introduces scaling prediction methods for performance of downstream tasks, enabling more precise model training configuration search
  - BitCPM -- Ultimate Ternary Quantization: Compresses model parameter bit-width to 3 values, achieving 90% extreme model bit-width reduction
  - Efficient Training Engineering Optimization: Adopts FP8 low-precision computing technology combined with Multi-token Prediction training strategy

- 📚 **High-Quality Training Data:**
  - UltraClean -- High-quality Pre-training Data Filtering and Generation: Builds iterative data cleaning strategies based on efficient data verification, open-sourcing high-quality Chinese and English pre-training dataset [UltraFinweb](https://huggingface.co/datasets/openbmb/Ultra-FineWeb)
  - UltraChat v2 -- High-quality Supervised Fine-tuning Data Generation: Constructs large-scale high-quality supervised fine-tuning datasets covering multiple dimensions including knowledge-intensive data, reasoning-intensive data, instruction-following data, and tool calling data

- ⚡ **Efficient Inference System:**
  - CPM.cu -- Lightweight and Efficient CUDA Inference Framework: Integrates sparse attention, model quantization, and speculative sampling to achieve efficient prefilling and decoding
  - ArkInfer -- Cross-platform Deployment System: Supports efficient deployment across multiple backend environments, providing flexible cross-platform adaptation capabilities

## Usage
### Inference with Transformers
```python
from transformers import AutoModelForCausalLM, AutoTokenizer
import torch

path = "openbmb/MiniCPM4-0.5B-QAT-Int4-unquantized"
device = "cuda"

tokenizer = AutoTokenizer.from_pretrained(path, trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained(path, torch_dtype=torch.bfloat16, device_map=device, trust_remote_code=True)

messages = [
    {"role": "user", "content": "推荐5个北京的景点。"},
]
model_inputs = tokenizer.apply_chat_template(messages, return_tensors="pt", add_generation_prompt=True).to(device)

model_outputs = model.generate(
    model_inputs,
    max_new_tokens=1024,
    top_p=0.7,
    temperature=0.7
)

output_token_ids = [
    model_outputs[i][len(model_inputs[i]):] for i in range(len(model_inputs))
]

responses = tokenizer.batch_decode(output_token_ids, skip_special_tokens=True)[0]
print(responses)

```

### Inference with [vLLM](https://github.com/vllm-project/vllm)

You can inference MiniCPM4-0.5B-QAT-Int4-unquantized with vLLM:
```python
from transformers import AutoTokenizer
from vllm import LLM, SamplingParams

model_name = "openbmb/MiniCPM4-0.5B-QAT-Int4-unquantized"
prompt = [{"role": "user", "content": "推荐5个北京的景点。"}]

tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True)
input_text = tokenizer.apply_chat_template(prompt, tokenize=False, add_generation_prompt=True)

llm = LLM(
    model=model_name,
    trust_remote_code=True,
    max_num_batched_tokens=32768,
    dtype="bfloat16", 
    gpu_memory_utilization=0.8, 
)
sampling_params = SamplingParams(top_p=0.7, temperature=0.7, max_tokens=1024, repetition_penalty=1.02)

outputs = llm.generate(prompts=input_text, sampling_params=sampling_params)

print(outputs[0].outputs[0].text)
```

## Evaluation Results
| Model          | Qwen3 | Llama3.2 | Gemma3 | MiniCPM4 | MiniCPM4 | MiniCPM4 |
|----------------|-------|----------|--------|----------|----------|----------|
| #Paramete      | 0.6B  | 1B       | 1B     | 0.5B     | 0.5B     | 0.5B     |
| #Precision     | BF16  | BF16     | BF16   | BF16     |Int4(Fake)|Int4(GPTQ)|
| MMLU           | 42.95 | 46.89    | 41.64  | 55.55    | 55.46    | 53.93    |
| CMMLU          | 42.05 | 23.73    | 25.09  | 65.22    | 63.91    | 63.73    |
| Ceval          | 45.53 | 36.74    | 31.83  | 66.11    | 64.85    | 65.22    |
| BBH            | 28.32 | 25.42    | 33.21  | 49.87    | 48.81    | 49.09    |
| GSM8K          | 61.71 | 39.76    | 61.26  | 52.08    | 45.41    | 45.49    |
| MBPP           | 47.86 | 47.47    | 59.92  | 59.14    | 55.64    | 55.25    |
| AVERAGE        | 44.73 | 36.66    | 42.15  | 58.00    | 55.68    | 55.45    |

## 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](https://github.com/OpenBMB/MiniCPM/blob/main/LICENSE) License. 

## Citation
- Please cite our [paper](https://github.com/OpenBMB/MiniCPM/tree/main/report/MiniCPM_4_Technical_Report.pdf) if you find our work valuable.

```bibtex
@article{minicpm4,
  title={{MiniCPM4}: Ultra-Efficient LLMs on End Devices},
  author={MiniCPM Team},
  year={2025}
}
```