RunLocalAI
gemma
31B parameters
Commercial OK
Multimodal

Gemma 4 31B Dense

Google's flagship dense Gemma 4. Beats some 400B-class proprietary models on benchmarks. Targets the 24GB single-GPU sweet spot.

License: Gemma Terms of Use·Released Apr 2, 2026·Context: 131,072 tokens

Overview

Google's flagship dense Gemma 4. Beats some 400B-class proprietary models on benchmarks. Targets the 24GB single-GPU sweet spot.

Strengths

  • Top dense ~30B model
  • Strong multilingual
  • 128K context

Weaknesses

  • Gemma license has use restrictions

Quantization variants

Each quantization trades model quality for file size and VRAM. Q4_K_M is the most popular starting point.

QuantizationFile sizeVRAM required
Q4_K_M18.0 GB24 GB
Q8_033.0 GB38 GB

Get the model

Ollama

One-line install

ollama run gemma4:31bRead our Ollama review →

HuggingFace

Original weights

huggingface.co/google/gemma-4-31b-it

Source repository — direct quantization required.

Hardware that runs this

Cards with enough VRAM for at least one quantization of Gemma 4 31B Dense.

NVIDIA GB200 NVL72
13824GB · nvidia
AMD Instinct MI355X
288GB · amd
AMD Instinct MI325X
256GB · amd
AMD Instinct MI300X
192GB · amd
NVIDIA B200
192GB · nvidia
NVIDIA H100 NVL
188GB · nvidia
NVIDIA H200
141GB · nvidia
Intel Gaudi 3
128GB · intel
Compare alternatives

Models worth comparing

Same parameter band, plus what's one tier above and below — so you can decide what actually fits your hardware.

Frequently asked

What's the minimum VRAM to run Gemma 4 31B Dense?

24GB of VRAM is enough to run Gemma 4 31B Dense at the Q4_K_M quantization (file size 18.0 GB). Higher-quality quantizations need more.

Can I use Gemma 4 31B Dense commercially?

Yes — Gemma 4 31B Dense ships under the Gemma Terms of Use, which permits commercial use. Always read the license text before deployment.

What's the context length of Gemma 4 31B Dense?

Gemma 4 31B Dense supports a context window of 131,072 tokens (about 131K).

How do I install Gemma 4 31B Dense with Ollama?

Run `ollama pull gemma4:31b` to download, then `ollama run gemma4:31b` to start a chat session. The default quantization is Q4_K_M.

Does Gemma 4 31B Dense support images?

Yes — Gemma 4 31B Dense is multimodal and accepts text + vision inputs. Vision support requires a runner that handles its image-conditioning architecture.

Source: huggingface.co/google/gemma-4-31b-it

Reviewed by RunLocalAI Editorial. See our editorial policy for how we research and verify model claims.