Google logo

DiffusionGemma

Gemmav26B A4BCurrent
byGoogleGoogle(enterprise)
Released June 10, 2026
19
OpenTools Scorenormalized /100
OpenTools Score
Context256K tokens
CategoryLarge Language Model

About DiffusionGemma

DiffusionGemma is Google DeepMind’s experimental open-weights text-diffusion model based on Gemma 4 26B A4B. It uses discrete diffusion and parallel canvas denoising to trade some benchmark quality for much faster local generation on dedicated GPUs.

Capabilities

textvisioncodereasoninglocal inference

Input Modalities

textimage

Output Modalities

text

Technical Details

API Identifier
google/diffusiongemma-26b-a4b-it
Category
Large Language Model
Context Window
256,000 tokens

Tags

diffusiongemmaopen-weightslocal-inferencemultimodalmixture-of-experts

Benchmarks

Performance scores for DiffusionGemma across standard benchmarks.

MMLU Proofficial-google-model-card · Jun 2026
77.6%
GPQA Diamondofficial-google-model-card · Jun 2026
73.2%
LiveCodeBench v6official-google-model-card · Jun 2026
69.1%
MMMLUofficial-google-model-card · Jun 2026
81.5%
HLE no toolsofficial-google-model-card · Jun 2026
11%
Google logo
Googleenterprise

Organizing the world's information with AI

14 Tools6 MCP Servers9 ModelsFounded 1998Mountain View, CA
View full profile

Tools by Google

Other AI tools from the same organization.

AI Models by Google

Large language models from the same organization.

ModelContext WindowPrice (In / Out per M)
Gemini Omni 1.1 FlashCurrent1.0M$1.50 / $17.50
Gemini 3.7 FlashCurrent1.0M$0.75 / $3.75
Gemini 3.5 Flash-LiteCurrent1.0M$0.30 / $2.50
Gemini 3.6 FlashCurrent1.0M$1.50 / $7.50
Gemma 4 26B A4BCurrent262K-- / --
Gemma 4 26B A4BCurrent262K$0.08 / $0.35

MCP Servers by Google

Connect this tool to AI assistants via the Model Context Protocol.

Related News

Latest coverage and updates.