Google launches DiffusionGemma model, freeing generation from left‑to‑right constraints

Google announced the release of DiffusionGemma, a new AI generation model. The model departs from the conventional left‑to‑right processing approach. It enables more flexible and efficient creation of images and text. Researchers say the architecture can improve coherence across generated content. DiffusionGemma is built on diffusion techniques refined by Google’s AI teams. The company plans to integrate the model into its cloud AI services. Early tests show potential gains in speed and quality over prior models. Users can expect broader creative capabilities as the technology rolls out.