DiffusionGemma: 4x faster text generation

DeepMind Blog

Google's DiffusionGemma generates 256 tokens simultaneously using text diffusion, achieving up to 4x faster inference than autoregressive models. It runs 1000+ tokens/s on an H100 and fits within 18GB VRAM when quantized. The 26B MoE model activates only 3.8B parameters per forward pass.

DiffusionGemmaGoogleGemma

See 14 more sources

NVIDIA Accelerates Google DeepMind’s DiffusionGemma for Local AI9 days agoMichael Fukuyama

Run DiffusionGemma on NVIDIA for Developer-Ready, High-Throughput Text Generation9 days agoAnu Srivastava

DiffusionGemma: The Developer Guide9 days ago

DiffusionGemma is our new experimental open model with up to 4x faster output on dedicated GPUs....9 days agoGoogle DeepMind

RT @osanseviero: Introducing DiffusionGemma, our first exploration with open diffusion text...9 days agoHugging Face

RT @googlegemma: Meet DiffusionGemma! An experimental open model that explores a fast approach to...9 days agoDemis Hassabis

RT @mervenoyann: DiffusionGemma is out 🔥 it's compute-bound so 4x faster compared to other Gemma-4...9 days agoHugging Face

DiffusionGemma9 days agoSimon Willison

Google's DiffusionGemma generates 256 tokens in parallel and self-corrects as it goes8 days ago

Google's DiffusionGemma AI Hits 1,000 Tokens Per Second—And It's Free9 days agoJose Antonio Lanz

Google's latest DiffusionGemma open AI model comes with a 4x speed boost9 days agoRyan Whitwam

Google’s DiffusionGemma is 4x faster than its other Gemma models9 days agoFrederic Lardinois

DiffusionGemma: The Developer Guide- Google Developers Blog9 days agotevlonDiscuss

Qwen Who? DiffusionGemma running at 1,500 tk/s on a Digital Pregnancy Test.8 days agoPorespellarDiscuss

DeepMind Blog

Jun 10, 4:24 PM