Unsloth Releases NVFP4 Quantized Gemma 4 for Faster Inference

Unsloth has released the NVFP4 quantized version of Gemma 4, supporting image-text-to-text tasks. This release aims to significantly reduce VRAM requirements, allowing the 12B model to run on just 11GB of VRAM while enabling faster GPU inference.

2026-07-15 ~ 2026-07-15 · 2 related posts

• unsloth Releases Quantized Gemma 4 — unsloth · 2026-07-14
• Gemma 4 NVFP4 Quantization Released — kalyan_kpl · 2026-07-15