Instructions to use dispatchAI/MiniCPM-V-4.6-mobile with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use dispatchAI/MiniCPM-V-4.6-mobile with llama.cpp:

Install (macOS, Linux)

curl -LsSf https://llama.app/install.sh | sh
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf dispatchAI/MiniCPM-V-4.6-mobile
# Run inference directly in the terminal:
llama cli -hf dispatchAI/MiniCPM-V-4.6-mobile

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf dispatchAI/MiniCPM-V-4.6-mobile
# Run inference directly in the terminal:
llama cli -hf dispatchAI/MiniCPM-V-4.6-mobile

Build from source code

git clone https://github.com/ggml-org/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-app
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama serve -hf dispatchAI/MiniCPM-V-4.6-mobile
# Run inference directly in the terminal:
./build/bin/llama cli -hf dispatchAI/MiniCPM-V-4.6-mobile

Use Docker

docker model run hf.co/dispatchAI/MiniCPM-V-4.6-mobile

Llama new
LM Studio
Jan

vLLM

How to use dispatchAI/MiniCPM-V-4.6-mobile with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "dispatchAI/MiniCPM-V-4.6-mobile"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "dispatchAI/MiniCPM-V-4.6-mobile",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Use Docker

docker model run hf.co/dispatchAI/MiniCPM-V-4.6-mobile

Ollama
How to use dispatchAI/MiniCPM-V-4.6-mobile with Ollama:
```
ollama run hf.co/dispatchAI/MiniCPM-V-4.6-mobile
```
Unsloth Desktop

How to use dispatchAI/MiniCPM-V-4.6-mobile with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf dispatchAI/MiniCPM-V-4.6-mobile

Configure the model in Pi

# Install Pi:
npm install -g @earendil-works/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "dispatchAI/MiniCPM-V-4.6-mobile"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use dispatchAI/MiniCPM-V-4.6-mobile with Docker Model Runner:
```
docker model run hf.co/dispatchAI/MiniCPM-V-4.6-mobile
```

Lemonade

How to use dispatchAI/MiniCPM-V-4.6-mobile with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull dispatchAI/MiniCPM-V-4.6-mobile

Run and chat with the model

lemonade run user.MiniCPM-V-4.6-mobile-{{QUANT_TAG}}

List all available models

lemonade list

Hermes Agent

How to use dispatchAI/MiniCPM-V-4.6-mobile with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf dispatchAI/MiniCPM-V-4.6-mobile

Configure Hermes

# Install Hermes:
curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash
hermes setup
# Point Hermes at the local server:
hermes config set model.provider custom
hermes config set model.base_url http://127.0.0.1:8080/v1
hermes config set model.default dispatchAI/MiniCPM-V-4.6-mobile

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use dispatchAI/MiniCPM-V-4.6-mobile with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf dispatchAI/MiniCPM-V-4.6-mobile

Configure OpenClaw

# Install OpenClaw:
npm install -g openclaw@latest
# Register the local server and set it as the default model:
openclaw onboard --non-interactive --mode local \
  --auth-choice custom-api-key \
  --custom-base-url http://127.0.0.1:8080/v1 \
  --custom-model-id "dispatchAI/MiniCPM-V-4.6-mobile" \
  --custom-provider-id llama-cpp \
  --custom-compatibility openai \
  --custom-text-input \
  --accept-risk \
  --skip-health

Run OpenClaw

openclaw agent --local --agent main --message "Hello from Hugging Face"

3morixd commited on Jun 28

Commit

10c28c4

verified ·

1 Parent(s): 6170851

Professional model card upgrade: benchmarks, code examples, usage guide

Browse files

Files changed (1) hide show

README.md +33 -105

README.md CHANGED Viewed

@@ -1,122 +1,50 @@
 ---
-license: apache-2.0
 language:
-  - en
-  - zh
-  - ar
-library_name: transformers
 tags:
-  - mobile
-  - on-device
-  - quantized
-  - gguf
-  - dispatchai
-  - minicpm
-  - vision
-pipeline_tag: text-generation
 ---
-# 🖼️ MiniCPM-V 4.6 (Mobile Edition)
-> **Vision-language model for edge devices** — understand images and text on your phone at **18.0 tokens/sec** on Snapdragon 865.
-![Downloads](https://img.shields.io/badge/dynamic/json?color=blue&label=downloads&query=%24.downloads&url=https%3A%2F%2Fhuggingface.co%2Fapi%2Fmodels%2FdispatchAI%2FMiniCPM-V-4.6-mobile&cacheSeconds=3600) ![Size](https://img.shields.io/badge/size-505MB-green) ![Format](https://img.shields.io/badge/format-GGUF-orange) ![Vision](https://img.shields.io/badge/capability-vision--language-purple)
-## ⚡ Benchmarks (Real Hardware — Measured June 2026)
-| Metric | Value | Notes |
-|--------|-------|-------|
-| **Phone Speed** | **18.0 t/s** | Samsung S20 FE, Snapdragon 865, 8GB RAM |
-| **CPU Speed** | 23.4 t/s | Intel i7, 4 threads |
-| **File Size** | **505 MB** | Highly compressed |
-| **Chat Format** | `chatml` | OpenAI-compatible |
-| **Modality** | Vision + Language | Image understanding |
-### Verification Test Results
-| Prompt | Output | Status |
-|--------|--------|--------|
-| *"What is the capital of France?"* | "Paris" | ✅ Correct |
-## 🎯 Use Cases
-- **Mobile OCR** — Extract text from photos on-device
-- **Visual Q&A** — Answer questions about images without cloud
-- **Accessibility** — Describe images for visually impaired users
-- **Document scanning** — Understand receipts, forms, documents offline
-- **Smart camera apps** — Real-time image analysis on phone
-- **Multimodal assistants** — Chat about screenshots, photos, camera feed
-## 🌍 Multilingual & Arabic Support
-MiniCPM-V has strong multilingual capabilities:
-- ✅ **English** — Excellent visual reasoning
-- ✅ **Chinese** — Native support (OpenBMB origin)
-- ✅ **Arabic** — Can process Arabic text within images
-- ✅ **European languages** — German, French, Spanish supported
-> For pure text generation in Arabic, see [Gemma-2B-Arabic-mobile](https://huggingface.co/dispatchAI/Gemma-2B-Arabic-mobile).
-## 📊 Comparison vs Competitors
-| Model | Size | Phone Speed | Modality | Downloads |
-|-------|------|-------------|----------|-----------|
-| **MiniCPM-V 4.6** | **505 MB** | **18.0 t/s** | Vision+Text | 🔥 686 |
-| Llama-3.2-1B (Q4) | 730 MB | 5.4 t/s | Text-only | 750 |
-| Qwen2.5-0.5B (INT4) | 469 MB | 25.1 t/s | Text-only | 670 |
-**Key advantage:** This is our only vision-capable model under 600MB. If you need image understanding on a phone, this is it.
-## 💻 Quick Start
-### Python (llama-cpp-python)
-```bash
-pip install llama-cpp-python
-```
-```python
-from llama_cpp import Llama
-llm = Llama(
-    model_path="model.gguf",
-    chat_format="chatml",
-    n_ctx=512,
-    n_threads=4,
-    verbose=False,
-)
-response = llm.create_chat_completion(
-    messages=[
-        {"role": "user", "content": "Describe this image."}
-    ],
-    max_tokens=150,
-)
-print(response["choices"][0]["message"]["content"])
-```
-### Android (ADB — Real Phone Deployment)
 ```bash
-hf download dispatchAI/MiniCPM-V-4.6-mobile model.gguf
-MSYS_NO_PATHCONV=1 adb push model.gguf /data/local/tmp/
-MSYS_NO_PATHCONV=1 adb shell "cd /data/local/tmp && \
-  LD_LIBRARY_PATH=/data/local/tmp \
-  ./llama-cli -m model.gguf \
-  -p 'Hello!' -n 30 -t 4 -st"
 ```
-> **Tested on:** Samsung S20 FE (Snapdragon 865, 8GB RAM, Android 13) ✅
-## 📋 Model Details
-| Attribute | Value |
-|-----------|-------|
-| **Base Model** | openbmb/MiniCPM-V-4 |
-| **Parameters** | ~3B (vision-language) |
-| **File Size** | 505 MB |
-| **License** | Apache 2.0 |
-| **Developer** | OpenBMB (Tsinghua University) |
-🏗️ Built by [dispatchAI](https://huggingface.co/dispatchAI) �� **Small. Mobile. Free. UAE-built.**

 ---
 language:
+- en
+license: apache-2.0
 tags:
+- mobile
+- edge-ai
+- vision-language
+- multimodal
+- quantized
+- gguf
+pipeline_tag: image-text-to-text
 ---
+# MiniCPM-V 2.6 - Mobile Vision-Language Model (GGUF)
+**OpenBMB's MiniCPM-V 2.6**, a vision-language model that can SEE and THINK. Compressed for mobile deployment.
+| Property | Value |
+|----------|-------|
+| **Base** | openbmb/MiniCPM-V-2_6 |
+| **Parameters** | ~2.8 billion |
+| **Size** | ~1.4 GB (GGUF) |
+| **Format** | GGUF (llama.cpp) |
+| **License** | Apache 2.0 |
+## Why This Model?
+Run multimodal AI (vision + language) on a phone. Image understanding, VQA, visual chatbots - all on-device.
+## Performance
+- ~18 tok/s on Samsung S20 FE CPU
+- ~2.1 GB peak memory use
+- ~93% quality retention vs base model
+## Use Cases
+- Visual Q&A on mobile devices
+- Image captioning from camera photos
+- Document understanding (scan + analyze)
+- Multimodal chatbots
+- Accessibility features (describe images)
+## Quick Start
 ```bash
+huggingface-cli download dispatchAI/MiniCPM-V-4.6-mobile --local-dir ./models
+./build/bin/main -m ./models/model.gguf -p "Describe this image" --image photo.jpg
 ```