FrontierThe story, in brief
Google's DiffusionGemma proves you don't need to train from scratch to build a text diffusion model
Google retrofitted Gemma into a diffusion model using <10% of training budget. Parallel generation hits 1.5K tokens/sec — but reasoning quality drops.
The KeyNews take
Why it matters
DiffusionGemma demonstrates a new path to faster inference by adapting existing models rather than training from scratch, with material speed gains but quality tradeoffs that matter for reasoning workloads.
The key facts
5 to knowDiffusionGemma adapted from Gemma 4 using <10% of original training budget
Generates 256 tokens in parallel (vs. one at a time in autoregressive)
Throughput: ~1,500 tokens per second
Quality trails original autoregressive model, especially on reasoning tasks
Model adaptation/retrofitting approach as alternative to from-scratch training
Go to the source
The Decoderthe-decoder.com
Publisher excerpt: Instead of training a new model from scratch, Google DeepMind retrofitted Gemma 4 into a diffusion model using less than 10 percent of the original training budget. DiffusionGemma generates 256 tokens in parallel instead of one at a time, hitting about 1,500 tokens per second. Quality still trails…