Black Forest Labs’ FLUX.2-small-decoder was introduced as an optimized image decoder that provides approximately 1.4 times faster performance and reduced VRAM usage compared to standard FLUX.2 decoders, beneficial especially at higher image resolutions for large-scale, production-grade image generation.
Google’s gemma-4-12B-it model offers unified multimodal understanding across text, image, and audio inputs using an encoder-free, single decoder transformer architecture. It delivers near the performance of much larger models while being compact enough to operate on 16GB of RAM, facilitating advanced multi-input AI workflows within AWS infrastructure.
Source & further reading
FLUX.2-small-decoder and gemma-4-12B-it models now available on Amazon SageMaker JumpStart ↗Source published 2026-08-10. This report was prepared with AI assistance from the linked primary-source material. Claims and limitations should be checked against the original announcement or research.