AI · AI model
Goku
Rectified flow foundation models for joint image and video generation
Goku is a family of visual foundation models developed by researchers from ByteDance and HKU, using rectified flow Transformers. It handles text-to-video, image-to-video, and text-to-image generation with top benchmark scores on VBench and GenEval.
What it's for
- Generate dynamic video sequences directly from descriptive text prompts
- Animate still images into continuous video clips
- Create photorealistic images from textual descriptions
Features
- Rectified flow Transformer architecture for joint image and video generation
- Text-to-video (T2V) generation with high motion smoothness and dynamic degree
- Image-to-video (I2V) capabilities
- Text-to-image (T2I) synthesis
- Top-ranking VBench score of 84.85 among competitive video models
- Carefully curated fine-grained image and video training data
Pricing
Open source. Free and open research project
Details
- Type
- AI model
- Category
- AI
- Topic
- AI Video Generation
- Open source
- Yes
- Platforms
- Self-hosted, Code library
- For
- AI researchers, machine learning engineers, and video generation developers
- Alternative to
- Runway Gen-3, Luma Dream Machine, Kling AI, CogVideoX, HunyuanVideo
- Tags
- video generation, text to video, image to video, bytedance, flow matching, diffusion
Why it was saved
- @EHuanglu: A free/open model covering video + voice + avatar lipsync in one, announced as a rival option for AI video generation.
Compare
Related
AI creative suite for generating and editing video, images, and animation
AI platform generating and editing video, image, and synchronized audio.
AI video and image generator known for cinematic motion and physics
AI-assisted multitrack video editor and text-to-video generator for creators and teams.
AI video and image generator with advanced motion control and physical simulation
Generative AI video model producing cinematic footage with native audio
Updated