---
title: MiniCPM-V-4-gguf
canonical_url: "https://www.modelscope.cn/models/OpenBMB/MiniCPM-V-4-gguf"
md_url: "https://www.modelscope.cn/models/OpenBMB/MiniCPM-V-4-gguf.md"
repository: OpenBMB/MiniCPM-V-4-gguf
last_updated: 2026-08-24
license: apache-2.0
pipeline_tag: image-text-to-text
tasks:
  - image-text-to-text
base_model:
  - openbmb/MiniCPM-V-4
base_model_relation: quantized
library_name:
  - pytorch
  - transformer
  - gguf
frameworks:
  - pytorch
language:
  - multilingual
downloads: 9588
stars: 4
tags:
  - minicpm-v
  - vision
  - ocr
  - multi-image
  - video
  - custom_code
  - gguf
---

# MiniCPM-V-4-gguf

> MiniCPM-V-4-gguf - OpenBMB 在 ModelScope 开源的模型。A GPT-4V Level MLLM for Single Image, Multi Image and Video on Your Phone

OpenBMB/MiniCPM-V-4-gguf 是 ModelScope 魔搭社区上的image-text-to-text模型，采用 apache-2.0 许可，基于 openbmb/MiniCPM-V-4 构建。

- **Repository**: OpenBMB/MiniCPM-V-4-gguf
- **License**: apache-2.0
- **Tasks**: image-text-to-text
- **Base model**: openbmb/MiniCPM-V-4
- **Tags**: minicpm-v, vision, ocr, multi-image, video, custom_code, gguf
- **Downloads**: 9588
- **Stars**: 4
- **Last updated**: 2026-08-24

Source: https://www.modelscope.cn/models/OpenBMB/MiniCPM-V-4-gguf

---

<h1>A GPT-4V Level MLLM for Single Image, Multi Image and Video on Your Phone</h1>

[GitHub](https://github.com/OpenBMB/MiniCPM-o) | [MiniCPM Wiki(Chinese)](https://modelbest.feishu.cn/wiki/UtWxwcERfiRIpIkBOjuc3h9tn1D) | [Demo](http://211.93.21.133:8889/)</a> 



## MiniCPM-V 4.0

**MiniCPM-V 4.0** is the latest efficient model in the MiniCPM-V series. The model is built based on SigLIP2-400M and MiniCPM4-3B with a total of 4.1B parameters. It inherits the strong single-image, multi-image and video understanding performance of MiniCPM-V 2.6 with largely improved efficiency. Notable features of MiniCPM-V 4.0 include:

- 🔥 **Leading Visual Capability.**
   With only 4.1B parameters, MiniCPM-V 4.0 achieves an average score of 69.0 on OpenCompass, a comprehensive evaluation of 8 popular benchmarks, **outperforming GPT-4.1-mini-20250414, MiniCPM-V 2.6 (8.1B params, OpenCompass 65.2) and Qwen2.5-VL-3B-Instruct (3.8B params, OpenCompass 64.5)**. It also shows good performance in multi-image understanding and video understanding.

- 🚀 **Superior Efficiency.**
  Designed for on-device deployment, MiniCPM-V 4.0 runs smoothly on end devices. For example, it devlivers **less than 2s first token delay and more than 17 token/s decoding on iPhone 16 Pro Max**, without heating problems. It also shows superior throughput under concurrent requests.

-  💫  **Easy Usage.**
  MiniCPM-V 4.0 can be easily used in various ways including **llama.cpp, Ollama, vLLM, SGLang, LLaMA-Factory and local web demo** etc. We also open-source iOS App that can run on iPhone and iPad. Get started easily with our well-structured [Cookbook](https://github.com/OpenSQZ/MiniCPM-V-CookBook), featuring detailed instructions and practical examples.


### Evaluation

<details>
<summary>Click to view single image results on OpenCompass. </summary>
<div align="center">
<table style="margin: 0px auto;">
    <thead>
        <tr>
            <th nowrap="nowrap" align="left">model</th>
            <th>Size</th>
            <th>Opencompass</th>
            <th>OCRBench</th>
            <th>MathVista</th>
            <th>HallusionBench</th>
            <th>MMMU</th>
            <th>MMVet</th>
            <th>MMBench V1.1</th>
            <th>MMStar</th>
            <th>AI2D</th>
        </tr>
    </thead>
    <tbody align="center">
        <tr>
            <td colspan="11" align="left"><strong>Proprietary</strong></td>
        </tr>
        <tr>
            <td nowrap="nowrap" align="left">GPT-4v-20240409</td>
            <td>-</td>
            <td>63.5</td>
            <td>656</td>
            <td>55.2</td>
            <td>43.9</td>
            <td>61.7</td>
            <td>67.5</td>
            <td>79.8</td>
            <td>56.0</td>
            <td>78.6</td>
        </tr>
        <tr>
            <td nowrap="nowrap" align="left">Gemini-1.5-Pro</td>
            <td>-</td>
            <td>64.5</td>
            <td>754</td>
            <td>58.3</td>
            <td>45.6</td>
            <td>60.6</td>
            <td>64.0</td>
            <td>73.9</td>
            <td>59.1</td>
            <td>79.1</td>
        </tr>
        <tr>
            <td nowrap="nowrap" align="left">GPT-4.1-mini-20250414</td>
            <td>-</td>
            <td>68.9</td>
            <td>840</td>
            <td>70.9</td>
            <td>49.3</td>
            <td>55.0</td>
            <td>74.3</td>
            <td>80.9</td>
            <td>60.9</td>
            <td>76.0</td>
        </tr>
        <tr>
            <td nowrap="nowrap" align="left">Claude 3.5 Sonnet-20241022</td>
            <td>-</td>
            <td>70.6</td>
            <td>798</td>
            <td>65.3</td>
            <td>55.5</td>
            <td>66.4</td>
            <td>70.1</td>
            <td>81.7</td>
            <td>65.1</td>
            <td>81.2</td>
        </tr>
        <tr>
            <td colspan="11" align="left"><strong>Open-source</strong></td>
        </tr>
        <tr>
            <td nowrap="nowrap" align="left">Qwen2.5-VL-3B-Instruct</td>
            <td>3.8B</td>
            <td>64.5</td>
            <td>828</td>
            <td>61.2</td>
            <td>46.6</td>
            <td>51.2</td>
            <td>60.0</td>
            <td>76.8</td>
            <td>56.3</td>
            <td>81.4</td>
        </tr>
        <tr>
            <td nowrap="nowrap" align="left">InternVL2.5-4B</td>
            <td>3.7B</td>
            <td>65.1</td>
            <td>820</td>
            <td>60.8</td>
            <td>46.6</td>
            <td>51.8</td>
            <td>61.5</td>
            <td>78.2</td>
            <td>58.7</td>
            <td>81.4</td>
        </tr>
        <tr>
            <td nowrap="nowrap" align="left">Qwen2.5-VL-7B-Instruct</td>
            <td>8.3B</td>
            <td>70.9</td>
            <td>888</td>
            <td>68.1</td>
            <td>51.9</td>
            <td>58.0</td>
            <td>69.7</td>
            <td>82.2</td>
            <td>64.1</td>
            <td>84.3</td>
        </tr>
        <tr>
            <td nowrap="nowrap" align="left">InternVL2.5-8B</td>
            <td>8.1B</td>
            <td>68.1</td>
            <td>821</td>
            <td>64.5</td>
            <td>49.0</td>
            <td>56.2</td>
            <td>62.8</td>
            <td>82.5</td>
            <td>63.2</td>
            <td>84.6</td>
        </tr>
        <tr>
            <td nowrap="nowrap" align="left">MiniCPM-V-2.6</td>
            <td>8.1B</td>
            <td>65.2</td>
            <td>852</td>
            <td>60.8</td>
            <td>48.1</td>
            <td>49.8</td>
            <td>60.0</td>
            <td>78.0</td>
            <td>57.5</td>
            <td>82.1</td>
        </tr>
        <tr>
            <td nowrap="nowrap" align="left">MiniCPM-o-2.6</td>
            <td>8.7B</td>
            <td>70.2</td>
            <td>889</td>
            <td>73.3</td>
            <td>51.1</td>
            <td>50.9</td>
            <td>67.2</td>
            <td>80.6</td>
            <td>63.3</td>
            <td>86.1</td>
        </tr>
        <tr>
            <td nowrap="nowrap" align="left">MiniCPM-V-4.0</td>
            <td>4.1B</td>
            <td>69.0</td>
            <td>894</td>
            <td>66.9</td>
            <td>50.8</td>
            <td>51.2</td>
            <td>68.0</td>
            <td>79.7</td>
            <td>62.8</td>
            <td>82.9</td>
        </tr>
    </tbody>
</table>
</div>

</details>

<details>
<summary>Click to view single image results on ChartQA, MME, RealWorldQA, TextVQA, DocVQA, MathVision, DynaMath, WeMath, Object HalBench and MM Halbench. </summary>

<div align="center">
<table style="margin: 0px auto;">
    <thead>
        <tr>
            <th nowrap="nowrap" align="left">model</th>
            <th>Size</th>
            <th>ChartQA</th>
            <th>MME</th>
            <th>RealWorldQA</th>
            <th>TextVQA</th>
            <th>DocVQA</th>
            <th>MathVision</th>
            <th>DynaMath</th>
            <th>WeMath</th>
            <th colspan="2">Obj Hal</th>
            <th colspan="2">MM Hal</th>
        </tr>
    </thead>
    <tbody>
        <tr>
            <td></td>
            <td></td>
            <td></td>
            <td></td>
            <td></td>
            <td></td>
            <td></td>
            <td></td>
            <td></td>
            <td></td>
            <td>CHAIRs↓</td>
            <td>CHAIRi↓</td>
            <td nowrap="nowrap">score avg@3↑</td>
            <td nowrap="nowrap">hall rate avg@3↓</td>
        </tr>
        <tbody align="center">
        <tr>
            <td colspan="14" align="left"><strong>Proprietary</strong></td>
        </tr>
        <tr>
            <td nowrap="nowrap" align="left">GPT-4v-20240409</td>
            <td>-</td>
            <td>78.5</td>
            <td>1927</td>
            <td>61.4</td>
            <td>78.0</td>
            <td>88.4</td>
            <td>-</td>
            <td>-</td>
            <td>-</td>
            <td>-</td>
            <td>-</td>
            <td>-</td>
            <td>-</td>
        </tr>
        <tr>
            <td nowrap="nowrap" align="left">Gemini-1.5-Pro</td>
            <td>-</td>
            <td>87.2</td>
            <td>-</td>
            <td>67.5</td>
            <td>78.8</td>
            <td>93.1</td>
            <td>41.0</td>
            <td>31.5</td>
            <td>50.5</td>
            <td>-</td>
            <td>-</td>
            <td>-</td>
            <td>-</td>
        </tr>
        <tr>
            <td nowrap="nowrap" align="left">GPT-4.1-mini-20250414</td>
            <td>-</td>
            <td>-</td>
            <td>-</td>
            <td>-</td>
            <td>-</td>
            <td>-</td>
            <td>45.3</td>
            <td>47.7</td>
            <td>-</td>
            <td>-</td>
            <td>-</td>
            <td>-</td>
            <td>-</td>
        </tr>
        <tr>
            <td nowrap="nowrap" align="left">Claude 3.5 Sonnet-20241022</td>
            <td>-</td>
            <td>90.8</td>
            <td>-</td>
            <td>60.1</td>
            <td>74.1</td>
            <td>95.2</td>
            <td>35.6</td>
            <td>35.7</td>
            <td>44.0</td>
            <td>-</td>
            <td>-</td>
            <td>-</td>
            <td>-</td>
        </tr>
        <tr>
            <td colspan="14" align="left"><strong>Open-source</strong></td>
        </tr>
        <tr>
            <td nowrap="nowrap" align="left">Qwen2.5-VL-3B-Instruct</td>
            <td>3.8B</td>
            <td>84.0</td>
            <td>2157</td>
            <td>65.4</td>
            <td>79.3</td>
            <td>93.9</td>
            <td>21.9</td>
            <td>13.2</td>
            <td>22.9</td>
            <td>18.3</td>
            <td>10.8</td>
            <td>3.9 </td>
            <td>33.3 </td>
        </tr>
        <tr>
            <td nowrap="nowrap" align="left">InternVL2.5-4B</td>
            <td>3.7B</td>
            <td>84.0</td>
            <td>2338</td>
            <td>64.3</td>
            <td>76.8</td>
            <td>91.6</td>
            <td>18.4</td>
            <td>15.2</td>
            <td>21.2</td>
            <td>13.7</td>
            <td>8.7</td>
            <td>3.2 </td>
            <td>46.5 </td>
        </tr>
        <tr>
            <td nowrap="nowrap" align="left">Qwen2.5-VL-7B-Instruct</td>
            <td>8.3B</td>
            <td>87.3</td>
            <td>2347</td>
            <td>68.5</td>
            <td>84.9</td>
            <td>95.7</td>
            <td>25.4</td>
            <td>21.8</td>
            <td>36.2</td>
            <td>13.3</td>
            <td>7.9</td>
            <td>4.1 </td>
            <td>31.6 </td>
        </tr>
        <tr>
            <td nowrap="nowrap" align="left">InternVL2.5-8B</td>
            <td>8.1B</td>
            <td>84.8</td>
            <td>2344</td>
            <td>70.1</td>
            <td>79.1</td>
            <td>93.0</td>
            <td>17.0</td>
            <td>9.4</td>
            <td>23.5</td>
            <td>18.3</td>
            <td>11.6</td>
            <td>3.6 </td>
            <td>37.2</td>
        </tr>
        <tr>
            <td nowrap="nowrap" align="left">MiniCPM-V-2.6</td>
            <td>8.1B</td>
            <td>79.4</td>
            <td>2348</td>
            <td>65.0</td>
            <td>80.1</td>
            <td>90.8</td>
            <td>17.5</td>
            <td>9.0</td>
            <td>20.4</td>
            <td>7.3</td>
            <td>4.7</td>
            <td>4.0 </td>
            <td>29.9 </td>
        </tr>
        <tr>
            <td nowrap="nowrap" align="left">MiniCPM-o-2.6</td>
            <td>8.7B</td>
            <td>86.9</td>
            <td>2372</td>
            <td>68.1</td>
            <td>82.0</td>
            <td>93.5</td>
            <td>21.7</td>
            <td>10.4</td>
            <td>25.2</td>
            <td>6.3</td>
            <td>3.4</td>
            <td>4.1 </td>
            <td>31.3 </td>
        </tr>
        <tr>
            <td nowrap="nowrap" align="left">MiniCPM-V-4.0</td>
            <td>4.1B</td>
            <td>84.4</td>
            <td>2298</td>
            <td>68.5</td>
            <td>80.8</td>
            <td>92.9</td>
            <td>20.7</td>
            <td>14.2</td>
            <td>32.7</td>
            <td>6.3</td>
            <td>3.5</td>
            <td>4.1 </td>
            <td>29.2 </td>
        </tr>
    </tbody>
</table>
</div>

</details>

<details>
<summary>Click to view multi-image and video understanding results on Mantis, Blink and Video-MME. </summary>
<div align="center">
<table style="margin: 0px auto;">
    <thead>
        <tr>
            <th nowrap="nowrap" align="left">model</th>
            <th>Size</th>
            <th>Mantis</th>
            <th>Blink</th>
            <th nowrap="nowrap" colspan="2" >Video-MME</th>
        </tr>
    </thead>
    <tbody>
        <tr>
            <td></td>
            <td></td>
            <td></td>
            <td></td>
            <td>wo subs</td>
            <td>w subs</td>
        </tr>
        <tbody align="center">
        <tr>
            <td colspan="6" align="left"><strong>Proprietary</strong></td>
        </tr>
        <tr>
            <td nowrap="nowrap" align="left">GPT-4v-20240409</td>
            <td>-</td>
            <td>62.7</td>
            <td>54.6</td>
            <td>59.9</td>
            <td>63.3</td>
        </tr>
        <tr>
            <td nowrap="nowrap" align="left">Gemini-1.5-Pro</td>
            <td>-</td>
            <td>-</td>
            <td>59.1</td>
            <td>75.0</td>
            <td>81.3</td>
        </tr>
        <tr>
            <td nowrap="nowrap" align="left">GPT-4o-20240513</td>
            <td>-</td>
            <td>-</td>
            <td>68.0</td>
            <td>71.9</td>
            <td>77.2</td>
        </tr>
        <tr>
            <td colspan="6" align="left"><strong>Open-source</strong></td>
        </tr>
        <tr>
            <td nowrap="nowrap" align="left">Qwen2.5-VL-3B-Instruct</td>
            <td>3.8B</td>
            <td>-</td>
            <td>47.6</td>
            <td>61.5</td>
            <td>67.6</td>
        </tr>
        <tr>
            <td nowrap="nowrap" align="left">InternVL2.5-4B</td>
            <td>3.7B</td>
            <td>62.7</td>
            <td>50.8</td>
            <td>62.3</td>
            <td>63.6</td>
        </tr>
        <tr>
            <td nowrap="nowrap" align="left">Qwen2.5-VL-7B-Instruct</td>
            <td>8.3B</td>
            <td>-</td>
            <td>56.4</td>
            <td>65.1</td>
            <td>71.6</td>
        </tr>
        <tr>
            <td nowrap="nowrap" align="left">InternVL2.5-8B</td>
            <td>8.1B</td>
            <td>67.7</td>
            <td>54.8</td>
            <td>64.2</td>
            <td>66.9</td>
        </tr>
        <tr>
            <td nowrap="nowrap" align="left">MiniCPM-V-2.6</td>
            <td>8.1B</td>
            <td>69.1</td>
            <td>53.0</td>
            <td>60.9</td>
            <td>63.6</td>
        </tr>
        <tr>
            <td nowrap="nowrap" align="left">MiniCPM-o-2.6</td>
            <td>8.7B</td>
            <td>71.9</td>
            <td>56.7</td>
            <td>63.9</td>
            <td>69.6</td>
        </tr>
        <tr>
            <td nowrap="nowrap" align="left">MiniCPM-V-4.0</td>
            <td>4.1B</td>
            <td>71.4</td>
            <td>54.0</td>
            <td>61.2</td>
            <td>65.8</td>
        </tr>
    </tbody>
</table>
</div>

</details>

### Examples

<div style="display: flex; flex-direction: column; align-items: center;">
  <img src="https://raw.githubusercontent.com/openbmb/MiniCPM-o/main/assets/minicpmv4/minicpm-v-4-case.png" alt="math" style="margin-bottom: 5px;">
</div>

Run locally on iPhone 16 Pro Max with [iOS demo](https://github.com/OpenSQZ/MiniCPM-V-CookBook/blob/main/demo/ios_demo/ios.md).

<div align="center">
  <img src="https://raw.githubusercontent.com/openbmb/MiniCPM-o/main/assets/minicpmv4/iphone_en.gif" width="45%" style="display: inline-block; margin: 0 10px;"/>
  <img src="https://raw.githubusercontent.com/openbmb/MiniCPM-o/main/assets/minicpmv4/iphone_en_information_extraction.gif" width="45%" style="display: inline-block; margin: 0 10px;"/>
</div>

<div align="center">
  <img src="https://raw.githubusercontent.com/openbmb/MiniCPM-o/main/assets/minicpmv4/iphone_cn.gif" width="45%" style="display: inline-block; margin: 0 10px;"/>
  <img src="https://raw.githubusercontent.com/openbmb/MiniCPM-o/main/assets/minicpmv4/iphone_cn_funny_points.gif" width="45%" style="display: inline-block; margin: 0 10px;"/>
</div> 

## Usage

```python
from PIL import Image
import torch
from transformers import AutoModel, AutoTokenizer

model_path = 'openbmb/MiniCPM-V-4'
model = AutoModel.from_pretrained(model_path, trust_remote_code=True,
                                  # sdpa or flash_attention_2, no eager
                                  attn_implementation='sdpa', torch_dtype=torch.bfloat16)
model = model.eval().cuda()
tokenizer = AutoTokenizer.from_pretrained(
    model_path, trust_remote_code=True)



image = Image.open('./assets/single.png').convert('RGB')

# First round chat 
question = "What is the landform in the picture?"
msgs = [{'role': 'user', 'content': [image, question]}]

answer = model.chat(
    msgs=msgs,
    image=image,
    tokenizer=tokenizer
)
print(answer)


# Second round chat, pass history context of multi-turn conversation
msgs.append({"role": "assistant", "content": [answer]})
msgs.append({"role": "user", "content": [
            "What should I pay attention to when traveling here?"]})

answer = model.chat(
    msgs=msgs,
    image=None,
    tokenizer=tokenizer
)
print(answer)
```


## License
#### Model License
* The code in this repo is released under the [Apache-2.0](https://github.com/OpenBMB/MiniCPM/blob/main/LICENSE) License. 
* The usage of MiniCPM-V series model weights must strictly follow [MiniCPM Model License.md](https://github.com/OpenBMB/MiniCPM/blob/main/MiniCPM%20Model%20License.md).
* The models and weights of MiniCPM are completely free for academic research. After filling out a ["questionnaire"](https://modelbest.feishu.cn/share/base/form/shrcnpV5ZT9EJ6xYjh3Kx0J6v8g) for registration, MiniCPM-V 2.6 weights are also available for free commercial use.


#### Statement
* As an LMM, MiniCPM-V 4.0 generates contents by learning a large mount of multimodal corpora, but it cannot comprehend, express personal opinions or make value judgement. Anything generated by MiniCPM-V 4.0 does not represent the views and positions of the model developers
* We will not be liable for any problems arising from the use of the MinCPM-V models, including but not limited to data security issues, risk of public opinion, or any risks and problems arising from the misdirection, misuse, dissemination or misuse of the model.

## Key Techniques and Other Multimodal Projects

👏 Welcome to explore key techniques of MiniCPM-V 2.6 and other multimodal projects of our team:

[VisCPM](https://github.com/OpenBMB/VisCPM/tree/main) | [RLHF-V](https://github.com/RLHF-V/RLHF-V) | [LLaVA-UHD](https://github.com/thunlp/LLaVA-UHD)  | [RLAIF-V](https://github.com/RLHF-V/RLAIF-V)

## Citation

If you find our work helpful, please consider citing our papers 📝 and liking this project ❤️！

```bib
@article{yao2024minicpm,
  title={MiniCPM-V: A GPT-4V Level MLLM on Your Phone},
  author={Yao, Yuan and Yu, Tianyu and Zhang, Ao and Wang, Chongyi and Cui, Junbo and Zhu, Hongji and Cai, Tianchi and Li, Haoyu and Zhao, Weilin and He, Zhihui and others},
  journal={Nat Commun 16, 5509 (2025)},
  year={2025}
}
```