DiffusionGemma: 4x faster text generation
DiffusionGemma is an experimental open model that achieves up to 4x faster text generation on GPUs compared to typical LLMs, using a diffusion head to maximize speed and operating as a 26B MoE model with 3.8B parameters active during inference. It is optimized for speed-critical workflows, supports bi-directional attention, and offers intelligent self-correction, but its quality is lower than standard Gemma 4. It is accessible via Hugging Face and supports various hardware platforms for efficient serving.
More from Deepmind Google
Putting sign language AI into users’ hands
Introducing sign-language-to-text (SL2T), our breakthrough model powering new sign language features for Deaf and hard of hearing users.
Gemini Robotics 2 brings whole body intelligence to robots
Gemini Robotics ER 2: powering robotics with video understanding, task orchestration, and multi-robot collaboration
Gemini Robotics ER 2 helps robots reason, collaborate, and solve real-world tasks. It represents a step change in video understanding, tool orchestration, and multi-robot collaboration for robotic applications.
