Unverified50% confidenceFactExact time
In the video diffusion latent-space generation stage, the pipeline uses three models—Seedance 2.0, MiniMax H3 and Wan 2.1—to generate 720p/1080p temporal frames.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/18/2026
First Seen
Valid until: 1/16/2027
Sources
Related Entities
Related Claims
VerifiedPrompt Studio can generate structured MiniMax H3 video scripts.69% similarUnverifiedThe MiniMax H3 model adopts a temporal compression strategy of merging 4 frames into one, where a single latent token spans 4 frames of video.69% similarUnverified借助 MiniMax H3 模型与 ComfyUI 上下文循环节点，可以在本地设备上生成时长超过一分钟的长视频69% similarUnverifiedThe workflow uses Flux 2 Klein for static images and Minimax H3 for video generation.69% similarUnverifiedMiniMax H3 is a LoRA model that is only just over 100 MB in size and can turn any static image into a video that can be viewed with 360-degree orbiting.69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1062372API
curl https://kongchang.com/api/v1/knowledge/claims/1062372MCP
get_claim(id=1062372)