Home / Models / MiniCPM-V 4.7 35B-A3B

How to Run MiniCPM-V 4.7 35B-A3B Locally — Setup Guide (October 2026)

MiniCPM-V 4.7 35B-A3B is OpenBMB's vision-language model and the first mixture-of-experts model in the MiniCPM-V line: about 3B parameters active per token, Apache-2.0, handling single images, multiple images and video. OpenBMB publishes the GGUFs itself.

Numbers you can trust

We don't print guessed speeds. Every measured number on this page came from TurboLLM's own auto-tuner running MiniCPM-V 4.7 35B-A3B on real hardware — labeled as such. When you run it yourself, TurboLLM benchmarks it on your exact GPU and shows a VRAM-fit verdict before you load, and the real measured tokens/sec once it's running.

Does it fit your GPU?

35B total, about 3B active parameters, built on the Qwen3.5 MoE text graph: 40 layers, 30 linear-attention (recurrent, no growing KV cache) and 10 full-attention (2 KV heads, 256 head dimension), 256 experts with 8 routed per token, and a 262,144-token context. Vision runs through a separate projector file (mmproj) in the same repo.

MiniCPM-V 4.7 35B-A3B
OpenBMB · GGUF
MoE · ~3B active · hybrid linear attention · vision
vision
VRAMFit
16 GBIQ3_XS (14.48 GB) plus the 1.12 GB vision projector file is a tight fit · keep the context short, or nudge some experts to system RAM
24 GBQ4_K_M (21.17 GB) in VRAM plus the vision projector · 262K native context
32 GBQ6_K (28.51 GB) in VRAM plus the vision projector · 262K native context
48 GB+Q8_0 (36.9 GB) fully in VRAM plus the vision projector · 262K native context
openbmb/MiniCPM-V-4.7-35B-A3B-gguf
View on Hugging Face →

How to run it

  1. Install TurboLLM

    npx turbollm — no install step, works on Windows, macOS, and Linux. It detects your GPU and auto-provisions a matching llama-server build (no CUDA toolkit, no Python, no compiler).

  2. Get the model

    Paste openbmb/MiniCPM-V-4.7-35B-A3B-gguf into TurboLLM's in-app Hugging Face search — it lists every quant with a VRAM-fit verdict against your real free VRAM, and downloads are resumable and SHA-256 verified.

  3. Let auto-fit pick the setup

    Flip the auto-fit toggle and TurboLLM chooses the GPU/CPU layer split (and, for a Mixture-of-Experts model, the expert-offload split) for your card — the same manual flag-hunting every other guide walks you through by hand.

  4. Load it

    TurboLLM benchmarks the model on your exact GPU as it loads and shows the real measured tokens/sec — not a number copied from someone else's card.

  5. Use it

    Chat in the built-in UI, or point any OpenAI- or Anthropic-compatible tool (including Claude Code) at TurboLLM's local API — see the API docs.

One honest caveat

OpenBMB's claims on the model card are vendor-run only: a higher overall score than Qwen3.6 35B-A3B in less inference time in Instruct mode, and 13+ tok/s in 3 GB on an iPhone 17 Pro Max through a Lookahead Expert Scheduling scheme in the llama.cpp-omni fork. The phone figure is not a desktop llama.cpp result. This page ships without a numbers table. Vision needs the projector file from the same repo.

FAQ

Will MiniCPM-V 4.7 35B-A3B run on my GPU?

Check the VRAM table above for the tier closest to your card. If you're not sure which tier you're in, or want picks tailored to your exact hardware, use What can I run? — enter your VRAM (or Apple unified memory) and it suggests the model, quant, and offload split that fits, the same logic TurboLLM's auto-fit runs when you load a model.

Which quant should I use for MiniCPM-V 4.7 35B-A3B?

Pick the highest quant that fits your VRAM with room left over for the KV cache and compute buffer — not just the model weights. As a starting point, Q4_K_M is usually the sweet spot (small quality loss for a large size drop); step up to Q5_K_M/Q6_K/Q8_0 if you have headroom, or down to Q3_K_M/an IQ quant if you don't. TurboLLM shows a VRAM-fit verdict for every quant against your card's real free VRAM before you download — see Quantization explained for what the letters and numbers actually mean, or llama.cpp flags that actually matter for the other settings worth knowing.

Why does MiniCPM-V 4.7 35B-A3B suddenly get slow or stutter?

The usual cause is silent VRAM spill: once a model's weights plus its KV cache stop fitting in VRAM, the driver quietly offloads the overflow to system RAM over PCIe, and that path runs 5–10× slower with no error message — just a sudden slowdown. TurboLLM's auto-fit sizes the quant and GPU/CPU split against your card's real free VRAM (including KV-cache growth at your chosen context) specifically to avoid this, and shows a VRAM-fit verdict before you load rather than after it's already crawling. Full mechanics in VRAM spill explained.

Run it now

$ npx turbollm

One command detects your GPU, provisions the right engine, and opens the UI. New here? Start with Install & first run and Quantization explained. On different hardware, see RTX 3090, RTX 4090 or RTX 5080. Also see how to run it or how to run it.