Expired85% confidenceFactExact time
Gemini Omni能够直接以视频作为输入源，从连续帧序列中提取运动轨迹、速度变化、物体交互等信息来生成新视频
1
Sources
85%
Confidence
Medium-term (~90 days)
Relevance
5/29/2026
First Seen
Valid until: 8/27/2026(expired)
Sources
Gemini Omni: How Powerful Is AI Video Generation That Understands Physics?
twitterGeminiApp5/29/2026
Related Entities
Related Claims
UnverifiedGemini Omni supports a video remix function that can automatically re-arrange multiple clips from different scenes according to the narrative pace or style described by the user.83% similarUnverifiedGemini Omni以视频作为条件输入而非仅依赖文本，理论上能将原视频中的运动动力学作为强约束，减少物理漂移82% similarUnverifiedGemini Omni标志着AI视频生成从'看起来像'向'物理上对'的方向演进79% similarVerifiedGemini Omni's ability to simultaneously understand users' language instructions, understand video content, and execute editing operations indicates that multimodal large models have reached a practical level in the closed-loop capability of understanding and generation.77% similarUnverifiedThe video generation feature of Gemini 3.1 Pro supports the generation of 8-second cinematic short clips.77% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/180API
curl https://kongchang.com/api/v1/knowledge/claims/180MCP
get_claim(id=180)