Unverified50% confidenceFactExact time
The MiniMax H3 generation of models supports first-frame and last-frame control.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/30/2026
First Seen
Valid until: 1/28/2027
Sources
Related Entities
Related Claims
UnverifiedMinimax is a multimodal large model from ByteDance, and H3 represents its image generation version.80% 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.80% similarVerifiedMiniMax-H3 features two capabilities: image-to-video generation and reference-based generation.72% similarUnverifiedfal released the H3 Max video generation model, which is based on post-training optimization of MiniMax H3.72% similarUnverifiedMiniMax H3在图生视频任务中通过时序建模和身份保持模块，能完整保留原图中的人物形象与画风特征，不出现形象崩坏72% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1117879API
curl https://kongchang.com/api/v1/knowledge/claims/1117879MCP
get_claim(id=1117879)