概要
Sora Overview#
Sora is a text-to-video generation model released by OpenAI in 2024, representing a major breakthrough in AI video generation.
Core Capabilities#
- Long Video Generation - Up to 60 seconds of video
- Physical World Understanding - Understanding physics and causality
- Complex Scene Handling - Coherent narrative with multiple characters and scenes
- High Resolution Output - Supports 1080p HD video
Technical Features#
- Based on Diffusion Transformer architecture
- Spacetime Patches technology
- Support for various aspect ratios and resolutions
Application Scenarios#
- Film previews and concept design
- Advertising and marketing content
- Educational and training videos
- Creative artistic expression
よくある質問
Sora 目前处于有限发布阶段，仅向部分创作者和安全研究人员开放。OpenAI 尚未公布公开发布的具体时间表。
