Intel AI Playground Updates bring you rapid editing, advanced video generation and more

Intel AI Playground version 2.6.0 has arrived with numerous practical updates that make locally generated models more easily predictable. This release introduces new models, including GPT-OSS 20B, WAN 2.1 VACE, and FLUX. The new backend version control settings allow users to choose which frameworks to be released to install and run, allowing them to stay up to date between playground updates. Outstanding workflows are edited at prompts with flux context. This allows you to modify or combine images using plain language without manual masking. This means you can merge two shots into a single cyberpunk scene just by removing the microphone, animating the photos, or explaining the editing.

This update also extends video capabilities for systems with Intel Arc discrete GPUs, using WAN 2.1 VACE to add inter-image and video-to-video workflows. The memo suggests that it starts at about 49 frames with 512×512 pixels to manage VRAM and iteration times, warning that a four-stage LORA setting can reduce calculations with degrees of freedom of motion. Convenience improvements include drag and drop thumbnails, direct comfyui launches for node-level adjustments, pytorch 2.8-equipped creation and enhancement, quick access modes for standard and HD, and a clean workflow menu. Image folders were made easier to sort using Yyyymmdd naming. llama.cpp exposes tunable context sizes and Vulkan backends to improve performance. Users can try more AI features locally and run the entire workflow from central software.