Google's DiffusionGemma LLM Achieves 1000 Tokens/Sec with Diffusion Architecture

Google DeepMind has released DiffusionGemma, an open-weight LLM that utilizes a diffusion architecture for text generation, enabling significantly faster inference speeds compared to traditional autoregressive models. This new model can process up to 1,000 tokens per second on a single H100 GPU and requires only 18 GB of VRAM, making it efficient for single-GPU deployments. While it trades some accuracy for speed, it excels in tasks like code infilling and real-time applications, and also supports multi-modal inputs including images and video. AI

IMPACT Accelerates inference speeds and reduces VRAM requirements, potentially enabling new real-time applications and wider single-GPU deployments.

RANK_REASON New open-weight diffusion-based LLM release from Google DeepMind. [lever_c_demoted from frontier_release: ic=1 ai=1.0]

Read on dev.to — LLM tag →

• DiffusionGemma
• Gemma 4
• Google Cloud Vertex AI Model Garden
• Google DeepMind
• Hugging Face Transformers
• JAX
• NVIDIA NIM
• Stable Diffusion
• vLLM

• model release
• infra

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →