Hosted on MSN

I tried Google's new DiffusionGemma, and watching it generate text like an image is unlike any local LLM

Most local LLMs feel predictable now. You download a model, point a runtime at it, ask a question, and then watch text move across the screen one token at a time. The ...
techtimes

Google’s DiffusionGemma Generates Text 4x Faster: Diffusion Replaces Token-by-Token Output

Google DeepMind released DiffusionGemma on June 10, 2026, an experimental open-weights model that writes text using discrete diffusion rather than the token-by-token method behind GPT-style systems, ...