GE

gemma4 31b imatrix mtp

GemmaGGUF by pearsonkyleUpdated 2026-07-11
Download on Hugging Face
Parameters30.7B
Context256K
Layers4
Architecturegemma4-assistant
Smallest quant0.4 GB
Quick answer

gemma4 31b imatrix mtp needs about 17.1 GB of memory at IQ3_M (a 14.4 GB file) with an 8K context window, so it fits on a 24 GB GPU. This repository has 5 quants, from 0.4 GB to 21.3 GB (Q5_K_S). Longer context needs more: at 32K tokens the KV cache alone adds about 8.6 GB.

Gemma4 31B with multimodal and tool support

Gemma4 31B is a 30.7B‑parameter model with a 262144‑token context window. It supports text, image, and tool usage, and is available in several quantizations for local deployment.

Good at

  • Multimodal text+image
  • Coding and tool use
  • Long‑context reasoning

Limits

  • No native vision without mmproj
  • Requires large VRAM for higher‑bit quants
  • License: gemma

Best for

  • Local inference on high‑VRAM GPUs
  • Multimodal chat assistants
  • Tool‑enabled coding workflows

Who it’s for. Developers needing a large, multimodal LLM that can run locally with llama.cpp or Ollama. Ideal for building agents that use text, images, and external tools.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
—0.4 GB 3.0 GB
Fits Get file →
IQ2_M10.9 GB 13.6 GB
Too big Get file →
IQ3_M14.4 GB 17.1 GB
Too big Get file →
IQ4_XS16.7 GB 19.4 GB
Too big Get file →
Q5_K_S21.3 GB 24.0 GB
Too big Get file →

For 8‑GB VRAM, use IQ2_M (10.92 GB file, 13.6 GB memory at 8k context). For 12‑GB VRAM, IQ3_M (14.42 GB file, 17.1 GB memory). For 24‑GB VRAM, Q5_K_S (21.31 GB file, 24 GB memory) plus the 0.36 GB MTP drafter and 0.75 GB vision mmproj fit on a single 24‑GB card.

What GPU do you need for gemma4 31b imatrix mtp?

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: the best quant that fits is the main file (0.4 GB file, about 3.0 GB in use).
  • 6 GB VRAM: the best quant that fits is the main file (0.4 GB file, about 3.0 GB in use).
  • 8 GB VRAM: the best quant that fits is the main file (0.4 GB file, about 3.0 GB in use).
  • 12 GB VRAM: the best quant that fits is the main file (0.4 GB file, about 3.0 GB in use).
  • 16 GB VRAM: the best quant that fits is IQ2_M (10.9 GB file, about 13.6 GB in use).
  • 24 GB VRAM: the best quant that fits is Q5_K_S (21.3 GB file, about 24.0 GB in use).

Which gemma4 31b imatrix mtp quant should you choose?

IQ3_M (14.4 GB) is the middle option in this repository. A reasonable choice when a 4-bit file will not fit, mostly on larger models. Quality loss is noticeable but often acceptable.

If memory is tight, the smallest file is the first one in the table at 0.4 GB, with the quality trade-offs that come with fewer bits.

New to quant names? Read GGUF quantization explained.

Specifications

Architecture from the GGUF header

Architecturegemma4-assistant
Parameters30.7B
Layers4
Trained context262,144 tokens
KV cache (fp16)256 KB per token

Source

Quantized bypearsonkyle
FamilyGemma
Licensegemma
Downloads (HF)809
Last updated2026-07-11

Frequently asked questions about gemma4 31b imatrix mtp

How much VRAM does gemma4 31b imatrix mtp need?

About 17.1 GB at IQ3_M with an 8K context window. That is the 14.4 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 gemma4 31b imatrix mtp on an 8 GB GPU?

Yes, with the smaller files: the 0.4 GB file needs about 3.0 GB at 8K context. Larger quants in this repository do not fit in 8 GB.

What is the file size of gemma4 31b imatrix mtp GGUF?

It depends on the quant: file is 0.4 GB, IQ2_M is 10.9 GB, IQ3_M is 14.4 GB, IQ4_XS is 16.7 GB, Q5_K_S is 21.3 GB.

What context length does gemma4 31b imatrix mtp support?

The GGUF header lists a trained context length of 262,144 tokens (256K). The KV cache grows with the context you use: about 2.1 GB at 8K and 8.6 GB at 32K.

Where can I download gemma4 31b imatrix mtp GGUF?

The files are on Hugging Face in the pearsonkyle/gemma4-31b-imatrix-mtp-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 gemma4 31b imatrix mtp use?

Hugging Face lists the license as gemma. 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?

Use the IQ2_M quant, which requires 13.6 GB memory at 8k context. 8 GB VRAM is insufficient for full inference.

Does the model support image input?

Yes, pair any quant file with the 772 MB mmproj vision tower to enable text+image processing.

What is the context window size?

The model supports up to 262144 tokens, allowing very long documents to be processed.

Other GGUFs of gemma-4-31B-it

Similar models

Browse all →

More from pearsonkyle

All by pearsonkyle →

More Gemma models

All Gemma →

Other medium models

All medium models →