Instructions to use prithivMLmods/MMFineReason-4B-f32-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use prithivMLmods/MMFineReason-4B-f32-GGUF with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("image-text-to-text", model="prithivMLmods/MMFineReason-4B-f32-GGUF")
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
pipe(text=messages)

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("prithivMLmods/MMFineReason-4B-f32-GGUF", dtype="auto")

llama-cpp-python

How to use prithivMLmods/MMFineReason-4B-f32-GGUF with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="prithivMLmods/MMFineReason-4B-f32-GGUF",
	filename="MMFineReason-4B.IQ4_XS.gguf",
)

llm.create_chat_completion(
	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"
					}
				}
			]
		}
	]
)

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use prithivMLmods/MMFineReason-4B-f32-GGUF with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf prithivMLmods/MMFineReason-4B-f32-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf prithivMLmods/MMFineReason-4B-f32-GGUF:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf prithivMLmods/MMFineReason-4B-f32-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf prithivMLmods/MMFineReason-4B-f32-GGUF:Q4_K_M

Use pre-built binary

# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf prithivMLmods/MMFineReason-4B-f32-GGUF:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf prithivMLmods/MMFineReason-4B-f32-GGUF:Q4_K_M

Build from source code

git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf prithivMLmods/MMFineReason-4B-f32-GGUF:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf prithivMLmods/MMFineReason-4B-f32-GGUF:Q4_K_M

Use Docker

docker model run hf.co/prithivMLmods/MMFineReason-4B-f32-GGUF:Q4_K_M

LM Studio
Jan

vLLM

How to use prithivMLmods/MMFineReason-4B-f32-GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "prithivMLmods/MMFineReason-4B-f32-GGUF"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "prithivMLmods/MMFineReason-4B-f32-GGUF",
		"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/prithivMLmods/MMFineReason-4B-f32-GGUF:Q4_K_M

SGLang

How to use prithivMLmods/MMFineReason-4B-f32-GGUF with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "prithivMLmods/MMFineReason-4B-f32-GGUF" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "prithivMLmods/MMFineReason-4B-f32-GGUF",
		"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 images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "prithivMLmods/MMFineReason-4B-f32-GGUF" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "prithivMLmods/MMFineReason-4B-f32-GGUF",
		"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"
						}
					}
				]
			}
		]
	}'

Ollama
How to use prithivMLmods/MMFineReason-4B-f32-GGUF with Ollama:
```
ollama run hf.co/prithivMLmods/MMFineReason-4B-f32-GGUF:Q4_K_M
```

Unsloth Studio

How to use prithivMLmods/MMFineReason-4B-f32-GGUF with Unsloth Studio:

Install Unsloth Studio (macOS, Linux, WSL)

curl -fsSL https://unsloth.ai/install.sh | sh
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for prithivMLmods/MMFineReason-4B-f32-GGUF to start chatting

Install Unsloth Studio (Windows)

irm https://unsloth.ai/install.ps1 | iex
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for prithivMLmods/MMFineReason-4B-f32-GGUF to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for prithivMLmods/MMFineReason-4B-f32-GGUF to start chatting

How to use prithivMLmods/MMFineReason-4B-f32-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf prithivMLmods/MMFineReason-4B-f32-GGUF:Q4_K_M

Configure the model in Pi

# Install Pi:
npm install -g @mariozechner/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "prithivMLmods/MMFineReason-4B-f32-GGUF:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use prithivMLmods/MMFineReason-4B-f32-GGUF with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf prithivMLmods/MMFineReason-4B-f32-GGUF:Q4_K_M

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 prithivMLmods/MMFineReason-4B-f32-GGUF:Q4_K_M

Run Hermes

hermes

Docker Model Runner
How to use prithivMLmods/MMFineReason-4B-f32-GGUF with Docker Model Runner:
```
docker model run hf.co/prithivMLmods/MMFineReason-4B-f32-GGUF:Q4_K_M
```

Lemonade

How to use prithivMLmods/MMFineReason-4B-f32-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull prithivMLmods/MMFineReason-4B-f32-GGUF:Q4_K_M

Run and chat with the model

lemonade run user.MMFineReason-4B-f32-GGUF-Q4_K_M

List all available models

lemonade list

MMFineReason-4B-f32-GGUF

MMFineReason-4B from OpenDataArena is a 4B-parameter vision-language model fine-tuned from Qwen3-VL-4B-Instruct on the massive MMFineReason dataset (1.8M high-quality samples, 5.1B solution tokens with average 2,910-token long-form Chain-of-Thought reasoning covering mathematics 79.4%, science 13.8%, puzzles/games 4.6%, and OCR/general 2.2%), using a two-stage pipeline of supervised fine-tuning (SFT) on MMFineReason-1.8M-SFT followed by reinforcement learning (RL) with GSPO on MMFineReason-1.8M-RL to achieve state-of-the-art multimodal reasoning performance that surpasses the larger Qwen3-VL-8B-Thinking (73.9 vs 72.5 average score). Trained via a systematic three-stage data-centric pipeline—large-scale collection/standardization, CoT rationale generation from Qwen3-VL-235B-A22B-Thinking, and difficulty-aware filtering revealing a "less is more" effect where just 7% (123K samples) matches full dataset performance—it excels on STEM diagrams, visual puzzles, complex charts (90.8% CharXiv, 75.6% RealWorldQA), and math benchmarks like 83.4% DynaMath while boosting general capabilities through reasoning-focused composition. Part of the MMFineReason family (2B/4B/8B scales), this parameter-efficient model demonstrates how high-quality, visually-grounded CoT data closes the gap between open-source VLMs and proprietary systems, available for deployment via standard Transformers frameworks.

MMFineReason-4B [GGUF]

File Name	Quant Type	File Size	File Link
MMFineReason-4B.IQ4_XS.gguf	IQ4_XS	2.49 GB	Download
MMFineReason-4B.Q2_K.gguf	Q2_K	1.8 GB	Download
MMFineReason-4B.Q3_K_L.gguf	Q3_K_L	2.41 GB	Download
MMFineReason-4B.Q3_K_M.gguf	Q3_K_M	2.24 GB	Download
MMFineReason-4B.Q3_K_S.gguf	Q3_K_S	2.05 GB	Download
MMFineReason-4B.Q4_K_M.gguf	Q4_K_M	2.72 GB	Download
MMFineReason-4B.Q4_K_S.gguf	Q4_K_S	2.6 GB	Download
MMFineReason-4B.Q5_K_M.gguf	Q5_K_M	3.16 GB	Download
MMFineReason-4B.Q5_K_S.gguf	Q5_K_S	3.09 GB	Download
MMFineReason-4B.Q6_K.gguf	Q6_K	3.63 GB	Download
MMFineReason-4B.Q8_0.gguf	Q8_0	4.69 GB	Download
MMFineReason-4B.f16.gguf	F16	8.83 GB	Download
MMFineReason-4B.f32.gguf	F32	17.7 GB	Download
MMFineReason-4B.i1-IQ1_M.gguf	i1-IQ1_M	1.25 GB	Download
MMFineReason-4B.i1-IQ1_S.gguf	i1-IQ1_S	1.18 GB	Download
MMFineReason-4B.i1-IQ2_M.gguf	i1-IQ2_M	1.68 GB	Download
MMFineReason-4B.i1-IQ2_S.gguf	i1-IQ2_S	1.58 GB	Download
MMFineReason-4B.i1-IQ2_XS.gguf	i1-IQ2_XS	1.48 GB	Download
MMFineReason-4B.i1-IQ2_XXS.gguf	i1-IQ2_XXS	1.37 GB	Download
MMFineReason-4B.i1-IQ3_M.gguf	i1-IQ3_M	2.13 GB	Download
MMFineReason-4B.i1-IQ3_S.gguf	i1-IQ3_S	2.07 GB	Download
MMFineReason-4B.i1-IQ3_XS.gguf	i1-IQ3_XS	1.98 GB	Download
MMFineReason-4B.i1-IQ3_XXS.gguf	i1-IQ3_XXS	1.84 GB	Download
MMFineReason-4B.i1-IQ4_NL.gguf	i1-IQ4_NL	2.6 GB	Download
MMFineReason-4B.i1-IQ4_XS.gguf	i1-IQ4_XS	2.48 GB	Download
MMFineReason-4B.i1-Q2_K.gguf	i1-Q2_K	1.8 GB	Download
MMFineReason-4B.i1-Q2_K_S.gguf	i1-Q2_K_S	1.69 GB	Download
MMFineReason-4B.i1-Q3_K_L.gguf	i1-Q3_K_L	2.41 GB	Download
MMFineReason-4B.i1-Q3_K_M.gguf	i1-Q3_K_M	2.24 GB	Download
MMFineReason-4B.i1-Q3_K_S.gguf	i1-Q3_K_S	2.05 GB	Download
MMFineReason-4B.i1-Q4_0.gguf	i1-Q4_0	2.59 GB	Download
MMFineReason-4B.i1-Q4_1.gguf	i1-Q4_1	2.84 GB	Download
MMFineReason-4B.i1-Q4_K_M.gguf	i1-Q4_K_M	2.72 GB	Download
MMFineReason-4B.i1-Q4_K_S.gguf	i1-Q4_K_S	2.6 GB	Download
MMFineReason-4B.i1-Q5_K_M.gguf	i1-Q5_K_M	3.16 GB	Download
MMFineReason-4B.i1-Q5_K_S.gguf	i1-Q5_K_S	3.09 GB	Download
MMFineReason-4B.i1-Q6_K.gguf	i1-Q6_K	3.63 GB	Download
MMFineReason-4B.imatrix.gguf	imatrix	3.87 MB	Download
MMFineReason-4B.mmproj-Q8_0.gguf	mmproj-Q8_0	454 MB	Download
MMFineReason-4B.mmproj-f16.gguf	mmproj-f16	836 MB	Download
MMFineReason-4B.mmproj-f32.gguf	mmproj-f32	1.66 GB	Download

Quants Usage

(sorted by size, not necessarily quality. IQ-quants are often preferable over similar sized non-IQ quants)

Here is a handy graph by ikawrakow comparing some lower-quality quant types (lower is better):

Downloads last month: 807

GGUF

Model size

4B params

Architecture

qwen3vl

Hardware compatibility

1-bit

2-bit

3-bit

4-bit

5-bit

6-bit

8-bit

16-bit

32-bit

View +1 variant

Model tree for prithivMLmods/MMFineReason-4B-f32-GGUF

Base model

Qwen/Qwen3-VL-4B-Instruct

Finetuned

OpenDataArena/MMFineReason-4B

Quantized

(3)

this model