MI

Minicpm

MiniCPMGGUF by trithemiusUpdated 2026-10-15
Download on Hugging Face
Parameters34.7B
Context256K
Layers41
Architectureqwen35moe
Smallest quant15.5 GB
Quick answer

Minicpm needs about 16.7 GB of memory at its main file (a 15.5 GB file) with an 8K context window, so it fits on a 24 GB GPU. This repository has 2 quants, from 15.5 GB to 69.4 GB (F16). Longer context needs more: at 32K tokens the KV cache alone adds about 2.8 GB.

MiniCPM-V 4.7 35B-A3B Abliterated – a mixed‑precision GGUF conversion for vision‑language tasks

MiniCPM-V 4.7 35B-A3B Abliterated is a 34.7B‑parameter model with a 3B active Mixture‑of‑Experts per token. It is quantized with a mixed Q3_K / Q4_K_M strategy and includes vision, reasoning, and tools abilities. The model is released under Apache‑2.0.

Good at

  • Vision‑language inference
  • Reasoning tasks
  • Tool usage

Limits

  • Requires large VRAM (≈16.7 GB at 8 k context)
  • License is Apache‑2.0
  • Safety alignment is intentionally weakened

Best for

  • Visual question answering
  • Image description
  • Document understanding

Who it’s for. Developers needing a multimodal LLM for local inference with llama.cpp. The model is suitable for research and production environments that can provide the required GPU memory.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
—15.5 GB 16.7 GB
Too big Get file →
F1669.4 GB 70.6 GB
Too big Get file →

For systems with around 16 GB VRAM, use the "unknown" quant file (15.5 GB, ≈16.7 GB memory). For larger memory budgets, the F16 quant file (69.38 GB, ≈70.6 GB memory) offers higher precision.

What GPU do you need for Minicpm?

These figures add the file size, an fp16 KV cache at an 8K context window (worked out from this model’s own header) and about 0.5 GB for the runtime. They are estimates, so leave a little headroom, and expect to use less context or a smaller quant if you are right at the limit.

  • 4 GB VRAM: no quant fits at an 8K context.
  • 6 GB VRAM: no quant fits at an 8K context.
  • 8 GB VRAM: no quant fits at an 8K context.
  • 12 GB VRAM: no quant fits at an 8K context.
  • 16 GB VRAM: no quant fits at an 8K context.
  • 24 GB VRAM: the best quant that fits is the main file (15.5 GB file, about 16.7 GB in use).
  • 32 GB VRAM: the best quant that fits is the main file (15.5 GB file, about 16.7 GB in use).
  • 48 GB VRAM: the best quant that fits is the main file (15.5 GB file, about 16.7 GB in use).

Specifications

Architecture from the GGUF header

Architectureqwen35moe
Parameters34.7B (approximate, mixture of experts)
Active parameters3B
Layers41
Trained context262,144 tokens
KV cache (fp16)82 KB per token

Source

Quantized bytrithemius
FamilyMiniCPM
Licenseapache-2.0
Downloads (HF)5,646
Last updated2026-10-15

Frequently asked questions about Minicpm

How much VRAM does Minicpm need?

About 16.7 GB at its main file with an 8K context window. That is the 15.5 GB file, plus the KV cache and roughly 0.5 GB of runtime overhead. Smaller quants need less and longer context needs more.

Can I run Minicpm on an 8 GB GPU?

Not at an 8K context. The smallest file needs about 16.7 GB, so you would need a larger GPU or CPU offloading.

What is the file size of Minicpm GGUF?

It depends on the quant: file is 15.5 GB, F16 is 69.4 GB.

What context length does Minicpm support?

The GGUF header lists a trained context length of 262,144 tokens (256K). The KV cache grows with the context you use: about 0.7 GB at 8K and 2.8 GB at 32K.

Where can I download Minicpm GGUF?

The files are on Hugging Face in the trithemius/MiniCPM-V-4.7-35B-A3B-abliterated-gguf repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.

What license does Minicpm use?

Hugging Face lists the license as apache-2.0. Licenses can differ between the quantized files and the original model, so check the repository and the base model before commercial use.

Can I run this model on 8 GB VRAM?

No. The smallest quantized version requires about 16.7 GB of VRAM at 8 k context.

Does the model support vision tasks?

Yes. It includes a vision tower for visual question answering and image description.

Is the model safe for general use?

The safety alignment is intentionally weakened; users must handle potential harmful content responsibly.

Other GGUFs of MiniCPM-V-4.7-35B-A3B-abliterated

Similar models

Browse all →

More from trithemius

All by trithemius →

More MiniCPM models

All MiniCPM →

Other large models

All large models →