There has been a surge of new releases in the visual AI space this month. Developers are pushing the boundaries of what is possible with video and image generation.
MIT Han Lab introduced DC-VideoGen to make video creation much faster. This framework improves the speed of video diffusion models significantly. NVlabs also released LongLive, a tool that creates long videos in real time. It maintains visual consistency for up to 240 seconds and allows for interactive prompting.
For video editing, Decart AI launched Lucy Edit Dev. It is the first open-weight model that lets users edit videos using text guides. In the image sector, inclusionAI unveiled Ming UniVision. This model handles image understanding, generation, and editing all in one place. Nunchaku Tech also released Qwen-Image-Edit-2509, a powerful model designed specifically for high-quality image manipulation.
# DC-VideoGen: Revolutionizing Video Generation Efficiency Breakthrough in Video AI Performance MIT Han Lab has introduced DC-VideoGen, a groundbreaking post-training framework designed to dramatically accelerate video diffusion models. The project […]
LongLive is a framework from NVlabs that generates long videos in real time with interactive prompting. The system produces videos up to 240 seconds while maintaining visual consistency across frames. […]
# Ming UniVision: Unified Image AI with Continuous Visual Tokenization inclusionAI has unveiled Ming UniVision, a groundbreaking multimodal large language model (MLLM) that unifies image understanding, generation, and editing within […]
# Nunchaku Qwen Image Edit 2509: Advanced Image Editing Model Nunchaku Tech has released a state-of-the-art image editing model, Qwen-Image-Edit-2509, designed to provide efficient and high-performance image manipulation. The model […]
# Lucy Edit Dev: Revolutionizing Text-Guided Video Editing Open-Source AI Breakthrough in Video Manipulation Decart AI has launched Lucy Edit Dev, the first open-weight foundation model for text-guided video editing. […]