OpenBMBMiniCPMMiniCPMV4_7ForConditionalGeneration

MiniCPM V 4.7 1B — Hardware Requirements & GPU Compatibility

VisionFunctions

MiniCPM-V 4.7 1B is the smallest model in OpenBMB's MiniCPM-V 4.7 line, a vision-language model with roughly 1.5 billion parameters that handles single images, multiple images, and video. It is designed for phones, laptops, and other resource-limited devices, and the card says it substantially outperforms Qwen3.5 0.8B in both overall score and inference efficiency. At this size it runs on almost any hardware, including CPU-only machines, with a Q4 build needing around 1 GB of memory plus the vision encoder. The context window is 262,144 tokens. It is released under the Apache 2.0 license, which allows commercial use. Visual tokens are compressed early inside the vision encoder, and the model supports mixed 4x and 16x compression so you can favor detail or speed. It runs in llama.cpp, Ollama, vLLM, and SGLang, with GGUF, AWQ, GPTQ, and FP8 builds, and it has mobile support for iOS, Android, and HarmonyOS.

1.6B params · dense262K contextApache 2.0Oct 20268.8K downloads369 likes61.8K quant downloads

Your hardware

Verdicts below are for the reference RTX 4090. Set yours to personalise the page.

Detection reads your browser's GPU name with no data sent anywhere.

Runs great on a 24 GB GPU

Runs from 8 GB (Q4_K_M · 256K)Best on 24 GB (Q4_K_M · ~512 tok/s · 256K)Macs from 32 GB (~66 tok/s)

Apache 2.0

Reference GPU · RTX 4090 24 GB

Estimates, ±20% · how we estimate

Run MiniCPM V 4.7 1B

One recommended way. Every other method, OS and quant is in the step-by-step guide.

Other ways

each opens step-by-step

Run in cloud

Fits on RTX 3060 12GB (10 GB headroom) · Q4_K_M

Generation speed: ~183 tok/s; generation speed
Cost per 1M output tokens: $0.10; per 1M output tokens

Deployon Vast.aiBest optionfrom $0.069/hr

Compare GPUs →

Run MiniCPM V 4.7 1B

1.6 GB VRAM at 16K context · fits the RTX 4090 (22.4 GB free)

Works on macOS, Windows and Linux.

Install Ollama (Linux)
$curl -fsSL https://ollama.com/install.sh | sh
Install Ollama (macOS / Windows)
Download and run the app; it installs the ollama command.
Download Ollama ↗
Run Q4_K_M
$ollama run hf.co/openbmb/MiniCPM-V-4.7-1B-gguf:Q4_K_M
Raise the context window
Type it inside the chat. For the API server instead, start it with OLLAMA_CONTEXT_LENGTH=16384 ollama serve.
$/set parameter num_ctx 16384
A larger context needs more memory than the weights alone.
Use the API
Ollama serves its API on http://localhost:11434.

Works on macOS, Windows and Linux.

Install LM Studio
Desktop app for macOS (Apple silicon), Windows and Linux.
lmstudio.ai ↗
Download the model
In the app press Ctrl/Cmd + Shift + M and paste https://huggingface.co/openbmb/MiniCPM-V-4.7-1B-gguf, then pick Q4_K_M. Or use the CLI (run LM Studio once first):
$lms get https://huggingface.co/openbmb/MiniCPM-V-4.7-1B-gguf@Q4_K_M
Open on Hugging Face ↗
Load and chat
Open the model loader, select the model and set Context Length to 16384. Leave GPU offload at maximum if it fits.
Start the local server (optional)
$lms server start

Works on macOS, Windows and Linux.

Install llama.cpp (macOS)
MacPorts, Nix and conda-forge also work.
$brew install llama.cpp
Install options ↗
Install llama.cpp (Windows)
$winget install llama.cpp
Install llama.cpp (Linux)
Download the prebuilt llama-*-bin-ubuntu-* archive for your GPU backend (CPU, Vulkan) from the releases page, unpack it and run the binaries from that folder. Homebrew (brew install llama.cpp) and conda-forge also work on Linux.
llama.cpp releases ↗
Start the server
Downloads the model on first run. The chat UI and an OpenAI-compatible API are at http://localhost:8080.
$llama-server -hf openbmb/MiniCPM-V-4.7-1B-gguf:Q4_K_M -c 16384 -ngl 99 --port 8080

Works on macOS, Windows and Linux.

Option A: Docker Model Runner
Docker Desktop: Settings > AI > Enable Docker Model Runner. Docker Engine on Linux: install the docker-model-plugin package.
$docker model run hf.co/openbmb/MiniCPM-V-4.7-1B-gguf:Q4_K_M
Model Runner docs ↗
The quant tag after the repo is how Hugging Face cards write it; if your Docker version rejects it, drop the tag and pick the quant by repo.
Option B: llama.cpp server image (NVIDIA)
Needs the NVIDIA Container Toolkit. Models are cached in the llama-cache volume. Open http://localhost:8080.
$docker run --gpus all -p 8080:8080 -v llama-cache:/root/.cache ghcr.io/ggml-org/llama.cpp:server-cuda -hf openbmb/MiniCPM-V-4.7-1B-gguf:Q4_K_M -c 16384 -ngl 99 --host 0.0.0.0 --port 8080
Image tags (cuda13, vulkan, rocm) ↗
Cache path inside the image is not in the docs; remove the -v flag if the volume complains.

Linux only.

Install vLLM
Linux only; on Windows run it inside WSL2. NVIDIA GPU shown; AMD ROCm, Intel and CPU builds have their own install pages.
$uv venv --python 3.12 --seed
$source .venv/bin/activate
$uv pip install vllm --torch-backend=auto
vLLM installation ↗
Serve the model
OpenAI-compatible API at http://localhost:8000/v1.
$vllm serve openbmb/MiniCPM-V-4.7-1B --max-model-len 16384
Runs the original ~3 GB weights, not the Q4_K_M GGUF.

macOS only.

MLX supports only macOS, so follow these steps for that setup.

Install mlx-lm
$pip install mlx-lm
Chat in the terminal
$mlx_lm.chat --model mlx-community/MiniCPM-V-4.7-1B-8bit
Downloads the model on first run. The MLX repo has its own quantisation; it may not match the quant selected above.
Or start a local server
Listens on http://localhost:8080.
$mlx_lm.server --model mlx-community/MiniCPM-V-4.7-1B-8bit

Works on macOS, Windows and Linux.

Install Jan
jan.ai ↗
Find the model
Open the Hub tab, paste this model ID in Search, and choose Q4_K_M if offered.
$openbmb/MiniCPM-V-4.7-1B-gguf
Use the local API (optional)
Settings > Local API Server > Start Server. It listens on http://127.0.0.1:1337.

Rent GPUs hourly to run Q4_K_M right there. Pricing comes from partner marketplaces and varies.

Available GPU rentals for Q4_K_M, prioritized by value
GPU	Speed	$/1M tokens	Rent
RTX 3060 12GBCheapestCheapest10 GB free	~183 tok/s	$0.10	Deploy on Vast.aifrom $0.069/hr (opens in new tab)	10 GB free · ~183 tok/s · $0.10 / 1M tokens
RTX 3070 (8 GB)6 GB free	~228 tok/s	$0.12	Deploy on Vast.aifrom $0.097/hr (opens in new tab)	8 GB · 6 GB free · ~228 tok/s · $0.12 / 1M tokens
RTX A4000 (16 GB)14 GB free	~228 tok/s	$0.13	Deploy on Vast.aifrom $0.11/hr (opens in new tab)	16 GB · 14 GB free · ~228 tok/s · $0.13 / 1M tokens
RTX 5060 Ti 16GB14 GB free	~228 tok/s	$0.17	Deploy on Vast.aifrom $0.14/hr (opens in new tab)	14 GB free · ~228 tok/s · $0.17 / 1M tokens
RTX A5000 (24 GB)22 GB free	~390 tok/s	$0.11	Deploy on RunPodfrom $0.16/hr (opens in new tab)	24 GB · 22 GB free · ~390 tok/s · $0.11 / 1M tokens
RTX 3080 Ti (12 GB)10 GB free	~463 tok/s	$0.11	Deploy on RunPodfrom $0.18/hr (opens in new tab)	12 GB · 10 GB free · ~463 tok/s · $0.11 / 1M tokens
Tesla V100 PCIe 16GB14 GB free	~457 tok/s	$0.12	Deploy on RunPodfrom $0.19/hr (opens in new tab)	14 GB free · ~457 tok/s · $0.12 / 1M tokens
RTX 4000 Ada (20 GB)18 GB free	~183 tok/s	$0.30	Deploy on RunPodfrom $0.20/hr (opens in new tab)	20 GB · 18 GB free · ~183 tok/s · $0.30 / 1M tokens

Compare GPUs →

How Much VRAM Does MiniCPM V 4.7 1B Need?

Select a quantization level; the listed hardware adheres to it.

VRAM	+32K	+128K	File	Quality	On RTX 4090 (reference)
0.8 GB	1.5 GB	4 GB	0.4 GB	2-bit · i-quant extreme compression — Importance-weighted 2-bit, extreme compression — significant quality loss	Fits: 256K context · ~840 tok/s
0.8 GB	1.6 GB	4 GB	0.4 GB	2-bit · i-quant extra small — Importance-weighted 2-bit, extra small	Fits: 256K context · ~799 tok/s
0.8 GB	1.6 GB	4 GB	0.5 GB	2-bit · i-quant small — Importance-weighted 2-bit, small	Fits: 256K context · ~780 tok/s
0.9 GB	1.6 GB	4 GB	0.5 GB	2-bit · i-quant medium — Importance-weighted 2-bit, medium	Fits: 256K context · ~753 tok/s
0.9 GB	1.7 GB	4.1 GB	0.5 GB	3-bit · i-quant — Importance-weighted 3-bit	Fits: 256K context · ~690 tok/s
1.0 GB	1.7 GB	4.1 GB	0.5 GB	2-bit · small K-quant — 2-bit small K-quant	Fits: 256K context · ~676 tok/s
1.0 GB	1.7 GB	4.2 GB	0.6 GB	3-bit · i-quant extra small — Importance-weighted 3-bit, extra small	Fits: 256K context · ~662 tok/s
1.0 GB	1.8 GB	4.2 GB	0.6 GB	3-bit · i-quant small — Importance-weighted 3-bit, small	Fits: 256K context · ~649 tok/s
1.0 GB	1.8 GB	4.2 GB	0.5 GB	2-bit · K-quant — 2-bit quantization with K-quant improvements	Fits: 256K context · ~649 tok/s
1.0 GB	1.8 GB	4.2 GB	0.5 GB	3-bit small — 3-bit small quantization	Fits: 256K context · ~636 tok/s
1.1 GB	1.8 GB	4.2 GB	0.6 GB	3-bit · i-quant medium — Importance-weighted 3-bit, medium	Fits: 256K context · ~624 tok/s
1.1 GB	1.9 GB	4.3 GB	0.6 GB	3-bit medium — 3-bit medium quantization	Fits: 256K context · ~590 tok/s
1.1 GB	1.9 GB	4.3 GB	0.6 GB	4-bit · legacy — 4-bit legacy quantization	Fits: 256K context · ~580 tok/s
1.1 GB	1.9 GB	4.3 GB	0.6 GB	3-bit large — 3-bit large quantization	Fits: 256K context · ~570 tok/s
1.2 GB	1.9 GB	4.4 GB	0.6 GB	4-bit · i-quant compact — Importance-weighted 4-bit, compact	Fits: 256K context · ~551 tok/s
1.2 GB	2 GB	4.4 GB	0.7 GB	4-bit · legacy+ — 4-bit legacy quantization with offset	Fits: 256K context · ~537 tok/s
1.2 GB	2 GB	4.4 GB	0.6 GB	4-bit small — 4-bit small quantization	Fits: 256K context · ~537 tok/s
1.2 GB	2 GB	4.4 GB	0.7 GB	4-bit · i-quant non-linear — Importance-weighted 4-bit, non-linear	Fits: 256K context · ~537 tok/s
1.3 GB	2 GB	4.5 GB	0.7 GB	4-bit medium — 4-bit medium quantization — most popular sweet spot	Fits: 256K context · ~512 tok/s
1.3 GB	2.1 GB	4.5 GB	0.7 GB	5-bit · legacy — 5-bit legacy quantization	Fits: 256K context · ~496 tok/s
1.4 GB	2.2 GB	4.6 GB	0.8 GB	5-bit · legacy+ — 5-bit legacy quantization with offset	Fits: 256K context · ~461 tok/s
1.4 GB	2.2 GB	4.6 GB	0.7 GB	5-bit small — 5-bit small quantization	Fits: 256K context · ~461 tok/s
1.5 GB	2.2 GB	4.6 GB	0.8 GB	5-bit medium — 5-bit medium quantization — good quality/size tradeoff	Fits: 256K context · ~449 tok/s
1.6 GB	2.4 GB	4.8 GB	0.8 GB	6-bit, very good — 6-bit quantization, very good quality	Fits: 256K context · ~402 tok/s
1.9 GB	2.7 GB	5.1 GB	1.1 GB	8-bit, near-lossless — 8-bit quantization, near-lossless	Fits: 256K context · ~345 tok/s
3.5 GB	4.2 GB	6.6 GB	2 GB	16-bit · bfloat — Brain floating point 16 — preferred for training	Fits: 256K context · ~189 tok/s

Selected Q4_K_M

4-bit medium quantization

1.6GBat 16K context

Weights: 0.9 GB
KV cache: 0.4 GB
Overhead: 0.3 GB

Fits RTX 4090 (reference) · 22.4 GB free · ~512 tok/s

≥50 tok/s needs ≥98 GB/s · RTX 5090 ~910, RTX 3090 Ti ~512

Which Hardware Can Run MiniCPM V 4.7 1B?

Q4_K_M · 1.3 GB

MiniCPM V 4.7 1B (Q4_K_M) requires 1.3 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 2+ GB is recommended. Using the full 262K context window can add up to 6.4 GB, bringing total usage to 7.7 GB. 52 GPUs can run it, including NVIDIA GeForce RTX 5090, NVIDIA GeForce RTX 3090 Ti. 176 devices with unified memory can run MiniCPM V 4.7 1B, including NVIDIA DGX H100, NVIDIA DGX A100 640GB.

By VRAM tier · typical GPUs and Macs at Q4_K_M

Typical hardwareQuantVRAMtok/sContext that fits

8 GB · RTX 4060Q4_K_M · 1.6 GB · ~138 tok/s · 256K1.6~138256K8 GB tier →

12 GB · RTX 3060 12GBQ4_K_M · 1.6 GB · ~183 tok/s · 256K1.6~183256K12 GB tier →

16 GB · RTX 5060 Ti 16GBQ4_K_M · 1.6 GB · ~228 tok/s · 256K1.6~228256K16 GB tier →

24 GB · RTX 4090referenceQ4_K_M · 1.6 GB · ~512 tok/s · 256K1.6~512256K24 GB tier →

Mac 32 GB · Mac Mini M4Q4_K_M · 1.6 GB · ~66 tok/s · 256K1.6~66256KMacs →

Mac 64 GB · Mac Studio M4 MaxQ4_K_M · 1.6 GB · ~299 tok/s · 256K1.6~299256KMacs →

Alternatives to MiniCPM V 4.7 1B

MiniCPM family · 16 models →

Similar models

6tok/s on RTX 4090

Model	Score	VRAM	tok/s	Why it is similar
Granite 4.0 1BIBM	—	2.6 GB	~452	same size
LFM2 VL 1.6BLiquid AI	—	1.5 GB	~496	same size
Qwen Base Invoicev1.01 1.5BLaaP-ai	—	1.7 GB	~512	same size
OLMo 2 0425 1B InstructAllen AI	—	1.7 GB	~449	same size
OpenMath Nemotron 1.5BNVIDIA	—	1.7 GB	~512	same size
ReaderLM v2jinaai	—	1.7 GB	~512	same size

Specifications

Publisher: OpenBMB
Family: MiniCPM
Parameters: 1.6B
Architecture: MiniCPMV4_7ForConditionalGeneration
Context Length: 262,144 tokens
Vocabulary Size: 248,094
Release Date: 2026-10-06
License: Apache 2.0

Resources

Frequently Asked Questions

How much VRAM does MiniCPM V 4.7 1B need?

MiniCPM V 4.7 1B requires 1.3 GB of VRAM at Q4_K_M, or 3.5 GB at BF16. Full 262K context adds up to 6.4 GB (7.7 GB total).

VRAM = Weights + KV Cache + Overhead

Weights = 1.6B × 4.8 bits ÷ 8 = 0.9 GB

KV Cache + Overhead ≈ 0.4 GB (at 2K context + ~0.3 GB framework)

Fit ratings and hardware model lists check this model with room for a 16K-token context, which needs a little more memory.

KV Cache + Overhead ≈ 6.8 GB (at full 262K context)

VRAM usage by quantization

Q4_K_M

1.3 GB

Q4_K_M + full context

7.7 GB

Learn more about VRAM estimation →

What's the best quantization for MiniCPM V 4.7 1B?

For MiniCPM V 4.7 1B, Q4_K_M (1.3 GB) offers the best balance of quality and VRAM usage. Q5_0 (1.3 GB) provides better quality if you have the VRAM. The smallest option is IQ2_XXS at 0.8 GB.

VRAM requirement by quantization

IQ2_XXS

0.8 GB~53%

IQ3_XS

1.0 GB~73%

Q3_K_L

1.1 GB~86%

Q4_K_M ★

1.3 GB~89%

Q5_0

1.3 GB~90%

BF16

3.5 GB~100%

★ Recommended — best balance of quality and VRAM usage.

Learn more about quantization →

Can I run MiniCPM V 4.7 1B on a Mac?

Yes — Mac Mini M1 (8 GB) and 155 other Macs can run MiniCPM V 4.7 1B. Apple Silicon uses unified memory, so the model shares RAM with the system. At Q4_K_M you need at least 1.3 GB of usable unified memory (RAM minus macOS overhead).

Can I run MiniCPM V 4.7 1B locally?

Yes — MiniCPM V 4.7 1B can run locally on consumer hardware. Popular tools include Ollama, LM Studio, and llama.cpp.

How fast is MiniCPM V 4.7 1B?

At Q4_K_M, MiniCPM V 4.7 1B can reach ~3750 tok/s on AMD Instinct MI350X. On NVIDIA GeForce RTX 4090: ~512 tok/s. Speed depends mainly on GPU memory bandwidth. Real-world results typically within ±20%.

tok/s = (bandwidth GB/s ÷ model GB) × efficiency

Example: NVIDIA B200 → 8000 ÷ 1.28 × 0.65 = ~4063 tok/s

Estimated speed at Q4_K_M (1.3 GB)

NVIDIA B200

~4063 tok/s

NVIDIA GeForce RTX 4090

~512 tok/s

NVIDIA B300

~4063 tok/s

AMD Instinct MI350X

~3750 tok/s

Real-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.

Learn more about tok/s estimation →

What's the download size of MiniCPM V 4.7 1B?

At Q4_K_M, the download is about 0.93 GB. The full-precision BF16 version is 3.11 GB. The smallest option (IQ2_XXS) is 0.43 GB.

Which GPUs can run MiniCPM V 4.7 1B?

52 consumer GPUs can run MiniCPM V 4.7 1B at Q4_K_M. Top options include AMD Radeon RX 6700 XT, AMD Radeon RX 6800, AMD Radeon RX 6800 XT. 52 GPUs have plenty of headroom for comfortable inference.

Which devices can run MiniCPM V 4.7 1B?

176 devices with unified memory can run MiniCPM V 4.7 1B at Q4_K_M (1.3 GB), including AMD Ryzen AI 9 HX 370 (Strix Point) Laptop, ASUS Ascent GX10, Apple iPhone 17 Pro, Asus ROG Flow Z13 (2025, Ryzen AI Max+ 395, 128 GB). Apple Silicon Macs use unified memory shared between CPU and GPU, making them well-suited for local LLM inference.