Google launches new DiffusionGemma 26B model with up to 4x faster text generation, runs on just 18GB of VRAM
Google has been quite frequently releasing new models lately. Following yesterday's Gemini 3.5 Live Translate, built specifically for real-time voice ...









