MBZUAI Institute of Foundation Models (IFM) · Open World Model
PAN 2
PAN 2 is an open 29B-parameter diffusion transformer that simulates how a scene changes after a text action, outputting video up to 20 seconds at 704×1248 and 24 fps. It pairs an LLM planner for structured world-state reasoning with a video model trained from scratch. IFM self-reports 34.0% full-plan completion on 100 multi-step tasks, the highest of the models it tested, but MiniMax-H3 leads on single actions and human preference. Weights, training code and inference code are released under Apache 2.0.
Open the source ↗Similar Video & Animation releases
- PixVid: Back to Pixels, Lifting Latent Video Diffusion into Pixel Space
Yanke Song, Shuo Chen, Chris Rockwell, Ming-Hsuan Yang, Xun Huang et al. · Pixel-space video diffusion · Week 47, 2026
- pi-H3
Hansheng Chen, Shengqu Cai, Jiaming Song, Gordon Wetzstein, Xinchen Yan and Leonidas Guibas · Decoder-free few-step video · Week 47, 2026
- MPEG-Diffusion: Mixed-Resolution Video Generation
Dennis Menn, Tianpei Gu, Guande He, Aryan V S, Yanke Song, Nan Liu, Desai Xie, Xun Huang, Chenfeng Xu and Diana Marculescu · Mixed-resolution video diffusion · Week 47, 2026
- Open-H3-Regenerate-2K
NVlabs (SANA / SoL-Engine team) · Open video refiner · Week 47, 2026
- Reshoot 4D
itsmetamike · 4D video reshooting · Week 47, 2026
- Seedance 2.5 Camera Control (pill pipeline)
Sirio Berati · Code-driven AI camera control · Week 47, 2026
Published by Quanta AI Labs, an AI agency in India. Browse every week in AI News Weekly.