Model Detail
diffusiongemma-26B-A4B-it-8bit
—diffusiongemma-26B-A4B-it-8bit is a multimodal model with 26B parameters released by mlx-community. The model is registered under the image-text-to-text pipeline tag on Hugging Face, distributed under the permissive apache-2.0 license.
diffusiongemma-26B-A4B-it-8bit ships with 26B parameters. Total weight footprint is approximately 7.7 GB, which is the relevant figure when planning local-inference VRAM. The apache-2.0 license is permissive, allowing commercial deployment and derivative work without per-seat fees, though attribution requirements still apply.
diffusiongemma-26B-A4B-it-8bit is best fit for mixed text-and-image reasoning tasks such as document understanding. Treat this as a starting matrix rather than a benchmark verdict — the right deployment usually depends on the specific evaluation suite that mirrors your workload.
How Transparent is DiffusionGemma?
arXiv:2606.20560v1 Announce Type: cross Abstract: LLM reasoning transparency is a critical affordance for understanding model decisions, mitigating misuse and misalignment, and debugging surprising model behaviors. However, DiffusionGemma performs a larger fraction of its computation in a continuous
Neither Parallel Nor Sequential: How DiffusionGemma Actually Commits Tokens
arXiv:2606.14620v1 Announce Type: new Abstract: Open diffusion language models are marketed as parallel, non-autoregressive decoders, yet the order in which a shipped checkpoint actually commits its tokens is almost never measured. We instrument DiffusionGemma 26B, a masked discrete-diffusion mixtur
