One email each morning — yesterday's AI, sortedGet it in your inbox

METAL LAB

Luma Integrates MiniMax H3 Video Model into Luma Agents

Generates 15-second 2K video with stereo sound in one go, using text, image, video, and audio references

이미지: X — 미디어·생성AI

Summary

  • Luma announced it has integrated the MiniMax H3 model into its platform, Luma Agents
  • The model can generate up to 15 seconds of 2K resolution video together with native stereo sound
  • Users can guide the creative direction by inputting various references, including text, images, video, and audio
Video from the source
통합 모델
MiniMax H3
탑재 플랫폼
Luma Agents
최대 영상 길이
15초
해상도
2K
오디오
네이티브 스테레오 사운드
입력 레퍼런스
텍스트, 이미지, 영상, 오디오
발표일
2026-08-06 (X 게시 기준)

MiniMax H3 Integrated into Luma Agents

Luma announced via its official X account that the MiniMax H3 model is now available on its platform, Luma Agents. With this integration, users can access MiniMax H3's video generation capabilities directly within Luma Agents without having to switch between separate tools.

15-Second 2K Video with Native Stereo Sound

According to Luma, MiniMax H3 can generate video up to 15 seconds long at 2K resolution, with native stereo sound attached to the video. The feature was highlighted for allowing both video and audio to be produced simultaneously in a single generation step, without a separate process of editing and merging visuals and sound.

ItemDetails
Max Length15 seconds
Resolution2K
AudioNative stereo
Input ReferencesText, image, video, audio

Multi-Reference Support Spanning Text, Image, Video, and Audio

A notable aspect of this update is the variety of input options. In addition to text prompts, users can provide images, video, and audio files as references to precisely specify the look, motion, and sound of the output. In its announcement, Luma stated, "You direct the look. The motion. The sound," emphasizing that creators can directly determine the direction of the final output.

A Unified Creative Process in One Continuous Workflow

Luma explained that this integration provides "more creative range" and "one continuous workflow." This suggests that users can now build a single workflow within Luma Agents that spans from planning to video and sound generation, without going through separate external models or tools. This approach aligns with the trend among several generative AI platforms toward unified, agent-based creation. Integration strategies from other generative AI platforms have also been covered on metallab.ai.

Remaining Questions

This announcement is based solely on a single official post from Luma on X, and detailed information—such as the specific architecture, performance benchmarks, pricing, or the exact scope of the rollout for MiniMax H3—has not yet been disclosed. Actual user experience and performance differences compared to existing models are reportedly areas that require further confirmation.