Instructions to use OsaurusAI/MiniCPM-V-4.7-1B-JANG_4M with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use OsaurusAI/MiniCPM-V-4.7-1B-JANG_4M with MLX:

# Make sure mlx-vlm is installed
# pip install --upgrade mlx-vlm

from mlx_vlm import load, generate
from mlx_vlm.prompt_utils import apply_chat_template
from mlx_vlm.utils import load_config

# Load the model
model, processor = load("OsaurusAI/MiniCPM-V-4.7-1B-JANG_4M")
config = load_config("OsaurusAI/MiniCPM-V-4.7-1B-JANG_4M")

# Prepare input
image = ["http://images.cocodataset.org/val2017/000000039769.jpg"]
prompt = "Describe this image."

# Apply chat template
formatted_prompt = apply_chat_template(
    processor, config, prompt, num_images=1
)

# Generate output
output = generate(model, processor, formatted_prompt, image)
print(output)

Notebooks
Google Colab
Kaggle
Local Apps Settings
LM Studio

How to use OsaurusAI/MiniCPM-V-4.7-1B-JANG_4M with Pi:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "OsaurusAI/MiniCPM-V-4.7-1B-JANG_4M"

Configure the model in Pi

# Install Pi:
npm install -g @earendil-works/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "mlx-lm": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "OsaurusAI/MiniCPM-V-4.7-1B-JANG_4M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent

How to use OsaurusAI/MiniCPM-V-4.7-1B-JANG_4M with Hermes Agent:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "OsaurusAI/MiniCPM-V-4.7-1B-JANG_4M"

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 OsaurusAI/MiniCPM-V-4.7-1B-JANG_4M

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use OsaurusAI/MiniCPM-V-4.7-1B-JANG_4M with OpenClaw:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "OsaurusAI/MiniCPM-V-4.7-1B-JANG_4M"

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 "OsaurusAI/MiniCPM-V-4.7-1B-JANG_4M" \
  --custom-provider-id mlx-lm \
  --custom-compatibility openai \
  --custom-text-input \
  --accept-risk \
  --skip-health

Run OpenClaw

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

Runtime: Osaurus 0.25.21 or newer (osaurus.json: required_osaurus_version).

OsaurusAI/MiniCPM-V-4.7-1B-JANG_4M

A JANG bundle of openbmb/MiniCPM-V-4.7-1B: a dense 1B model (24 layers, GatedDeltaNet linear attention and full attention 3:1), with image, multi-image and video understanding, thinking on by default, and Qwen-style XML tool calls. 3.0 GiB of bf16 weights in 1.6 GiB.

What is quantized how

part	format
MLP gate / up	affine 4-bit, group 64, GPTQ
MLP down	affine 6-bit, group 64, GPTQ
attention q/k/v/o, GatedDeltaNet projections	affine 8-bit, GPTQ
embeddings, lm_head	affine 8-bit
vision tower + projector	affine 8-bit (fc2 bf16)
norms, conv, A_log, dt_bias	source values

Every quantized matrix with an input Hessian was rounded with GPTQ. The Hessians come from the model's own outputs: the bf16 model wrote reasoning traces (thinking on and off), tool calls, code, long-document answers and multi-turn agent turns across 11 prompt families at its native sampling settings, plus 192 images and videos through the vision path, and those sequences were the calibration data. Affine quantization is MLX's own grid (bf16 scales and biases), so the bundle loads at full speed on the standard kernels.

Fidelity vs the bf16 model

126,332 teacher-forced positions on 164 held-out sequences the bf16 model generated itself (never seen in calibration), scored against the bf16 model: top-128 KL, top-1 agreement, and the change in log-likelihood of the real next token (dNLL).

	size	KL median	KL mean	top-1	dNLL
bf16 model run in bf16 (the floor)	3.0 GiB	4.9e-06	0.0005	99.24%	+0.0003
1B-JANG_4M	1.6 GiB	3.5e-04	0.0270	95.27%	+0.0277

By prompt family (KL median, top-1):

family	KL median	top-1
agentic / tool use	0.0001	94.4%
coding	0.0085	92.2%
creative writing	0.0279	88.9%
cybersecurity	0.0182	90.6%
long documents	0.0059	93.1%
math	0.0000	97.1%
networking	0.0000	96.4%
office tasks	0.0000	97.2%
science	0.0008	95.0%
terminal agent	0.0009	95.1%
turn after a tool result	0.0003	94.5%

Decisions: the quantized model picks the same token as bf16 at 78 / 78 </think> closes and at 29 / 33 tool-call starts (bf16 vs itself: 29 / 33).

Model notes

Modalities: text, images (multi-image, high-resolution slicing), video. No audio. downsample_mode 16x (default) or 4x selects the visual token budget.
Reasoning: on by default; the chat template opens <think>. enable_thinking: false turns it off. There are no reasoning-effort levels.
Tools: Qwen3-Coder-style XML (<tool_call><function=…><parameter=…>).
Sampling (generation_config.json, mirrored in jang_config.json): temperature 1.0, top-p 0.95, top-k 20, EOS <|im_end|> / <|endoftext|>. The 1B release ships no sampling values; these are the 35B values.
The source config lists one MTP layer, but the checkpoint contains no MTP weights; none is included.

Source

openbmb/MiniCPM-V-4.7-1B revision 984a89aafcd5, Apache-2.0.
Quantization and bundle by Jinho Jang (eric@osaurus.ai), 2026-10-12.

Downloads last month: 52

Safetensors

Model size

2B params

Tensor type

U32

BF16

MLX

Hardware compatibility

8-bit

Model tree for OsaurusAI/MiniCPM-V-4.7-1B-JANG_4M

Base model

openbmb/MiniCPM-V-4.7-1B

Quantized

(53)

this model