Gemma 4 26B A4B

Open SourceMultimodal

About

Mixture of Experts (MoE) model with 26B total parameters, activating 4B per token. Efficient high-performance inference with advanced reasoning, supporting text, image, and video inputs.

Capabilities

VisionMultimodalReasoningFunction CallingTool UseJSON ModeCode Execution

Rankings

Best Open Source LLMs Best Multimodal / Vision LLMs Best LLMs for Function Calling & Tool Use Cheapest LLM APIs

Specifications

FamilyGemma 4

Released2026-04-02

Parameters26000000000

Context256k

Specializationmixture-of-experts