Unverified50% confidenceFactExact time
OpenAI excels in video generation (Sora) but has not yet invested equivalent resources in video understanding and reasoning.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
8/7/2026
First Seen
Valid until: 11/5/2026
Sources
为何只有Google做好了视频理解大模型？数据、架构与成本的三重壁垒
Related Claims
UnverifiedMainstream video generation models including Sora (OpenAI), Wan2.1 (Alibaba), and CogVideoX (Zhipu) have all adopted the DiT or its variant architecture.77% similarUnverified开源AI视频项目包括CogVideo、Open-Sora、Mochi等75% similarUnverifiedOpenAI's Sora and Google's VideoPoet both adopt the technical approach of compressing video frames into discrete tokens followed by autoregressive prediction73% similarVerifiedThe mainstream video generation models in the current AI video generation field include OpenAI's Sora, Kuaishou's Kling, Google's Veo, and Runway Gen-3.73% similarVerifiedOpenAI's Sora uses a diffusion Transformer architecture and is known for high-quality text-to-video generation.72% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/705522API
curl https://kongchang.com/api/v1/knowledge/claims/705522MCP
get_claim(id=705522)