Setup diffusiongemma-26B-A4B-it-NVFP4 PC with NPU

Setup diffusiongemma-26B-A4B-it-NVFP4 PC with NPU

Running this model locally is fastest when deployed through a PowerShell script.

Use the instructions provided below to complete the setup.

The loader auto-caches the model archive (several GBs included).

The installer diagnoses your environment to deploy the most compatible profile.

🔧 Digest: 8a993d4a20248899a53f570e602af59e • 🕒 Updated: 2026-06-28



  • Processor: 6-core 3.5 GHz minimum required
  • RAM: required: 16 GB absolute minimum for small models
  • Disk Space: at least 100 GB for multiple local LLM variants
  • GPU: 16 GB+ video memory highly recommended for exl2 / AWQ formats

The diffusiongemma-26B-A4B-it-NVFP4 model leverages a Gemma-based architecture to deliver high‑fidelity image generation with only 26 billion parameters. Its NVFP4 quantization enables fast inference on consumer‑grade hardware while preserving fine‑grained details. The model excels in multi‑modal prompting, accepting text instructions and producing corresponding visual outputs with impressive coherence. Compared to earlier diffusion models, it achieves a superior balance between speed and quality, making it suitable for real‑time creative workflows. Developers appreciate its seamless integration with the Transformer ecosystem and the built‑in support for conditional generation. Overall, the diffusiongemma-26B-A4B-it-NVFP4 stands out as a versatile tool for both research and production environments.

Parameter Count 26 B
Architecture Gemma‑based diffusion Transformer
Quantization NVFP4
Max Input Tokens 1024
Output Resolution 1024×1024