gemma
26B parameters
Commercial OK
Multimodal

Gemma 4 26B MoE

MoE variant of Gemma 4. Faster per-token than the 31B dense at similar quality on most tasks.

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

Overview

MoE variant of Gemma 4. Faster per-token than the 31B dense at similar quality on most tasks.

Strengths

  • MoE speed advantage
  • Multilingual
  • Multimodal

Weaknesses

  • MoE routing variance

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_M16.0 GB20 GB

Get the model

Ollama

One-line install

ollama run gemma4:26b-moeRead our Ollama review →

HuggingFace

Original weights

huggingface.co/google/gemma-4-26b-moe-it

Source repository — direct quantization required.

Hardware that runs this

Cards with enough VRAM for at least one quantization of Gemma 4 26B MoE.

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 26B MoE?

20GB of VRAM is enough to run Gemma 4 26B MoE at the Q4_K_M quantization (file size 16.0 GB). Higher-quality quantizations need more.

Can I use Gemma 4 26B MoE commercially?

Yes — Gemma 4 26B MoE 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 26B MoE?

Gemma 4 26B MoE supports a context window of 131,072 tokens (about 131K).

How do I install Gemma 4 26B MoE with Ollama?

Run `ollama pull gemma4:26b-moe` to download, then `ollama run gemma4:26b-moe` to start a chat session. The default quantization is Q4_K_M.

Does Gemma 4 26B MoE support images?

Yes — Gemma 4 26B MoE is multimodal and accepts text + vision inputs. Vision support requires a runner that handles its image-conditioning architecture.

Source: huggingface.co/google/gemma-4-26b-moe-it

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