Gemma 4 QAT models: Optimizing model compression for mobile and laptop efficiency

2026-06-08

Google released Gemma 4 checkpoints optimized with Quantization-Aware Training (QAT) to enhance efficiency on mobile and laptops. QAT minimizes performance loss during model compression, enabling models to run on everyday edge devices. The update includes a specialized mobile quantization format, significantly reducing memory requirements while maintaining model quality.