Beijing Lab Open-Sources 114B MoE Audio/Video Model MAGI-2
ShanRizvi · x · 2026-08-15
A new Beijing-based lab founded by Swin Transformer creators Cao Yue and Zhang Zheng has open-sourced MAGI-2 Preview. This is a 114B parameter Mixture-of-Experts (MoE) model for audio and video generation. It activates only 6B parameters per token via ultra-fine-grained head routing and custom MagiMoE kernels, supporting single-stream native A/V generation at practical scale. The generation cost is claimed to be 1/10th of mainstream models. Code and tech report are fully available.
Related event: SandAI Open-Sources MAGI-2, a 114B-Parameter MoE Video Model(4 posts)→
More from Multimodal
- Unreleased 'Odyssey' Demo Shows AI's True Potential — shauseth · 2026-08-15
- Grimfel launches free AI video generation tool — VoidStateKate · 2026-08-15
- Adobe Firefly transforms still images into moving memory — azed_ai · 2026-08-15
- Minimax H3 Video Gen Test: 5 Minutes vs Canva’s 30 — techhalla · 2026-08-15
- LichtFeld Studio: All-in-one tool for 3D Gaussian Splatting — tom_doerr · 2026-08-15
- Minimax H3 video generation shows horizontal tiling seam artifact, users seek fix — MrGrindor · 2026-08-15