Skip to content
On AI Radar
العربية

New Model

New ModelHugging Face ·

Meta releases Muse Spark Open — ~816B multimodal MoE, one million tokens

Meta published Muse Spark Open, an open-weight multimodal Mixture-of-Experts model for text, images, audio and video. It has ~816B parameters, ~42B active per token and a one million tokens context window.

Why it matters

Developers can run agentic coding workflows that read a codebase, write and debug code, and call tools on multimodal inputs across long contexts.

huggingface.co ↗
New ModelHugging Face ·

Black Forest Labs FLUX.3 Image Klein added to Diffusers

FLUX.3 Image Klein is available in Diffusers with text-to-image, multi-reference image-to-image and bounding-box layout control; distilled and base checkpoints are published on Hugging Face.

Why it matters

Developers and creators can iterate faster and control composition precisely: the distilled checkpoint makes a 1024 × 1024 image in about 2 seconds on an H100, speeding prototyping.

Worth tryinghuggingface.co ↗