freym
freym / models prices / MiniMax H3

MiniMax H3

MiniMax H3 is MiniMax's flagship model. Through late July and early August 2026 it spread across creative platforms — available on Leonardo, in Luma Agents, and via Pika's API Club.

By MiniMax · large language model · Last update tracked August 6, 2026 · Page updated 2026-08-08

Key facts

DeveloperMiniMax
Categorylarge language model
Notable forAvailable on Leonardo, Luma Agents and Pika API Club.
Available viaLuma AI, Pika, Leonardo
Tracked announcements3

Announcement timeline

availability

MiniMax H3 now available in Luma Agents

MiniMax H3 now available in Luma Agents

MiniMax H3 video generation is now integrated into Luma Agents, enabling up to 15 seconds of 2K video with native stereo sound, guided by text, image, video, and audio references in a unified workflow.

availability

H3 model now available via Pika API Club

MiniMax's H3 model is now accessible through Pika API Club at competitive pricing, up to 50% cheaper than other API aggregators.

availability

MiniMax H3 now available on Leonardo

MiniMax H3 now available on Leonardo

MiniMax H3, a video model that generates both video clips and soundtracks with locked character and voice references, is now available on Leonardo. The model offers commercial-grade quality for ads, e-commerce, gaming and UI with best-in-class value.

FAQ

What is MiniMax H3?

MiniMax H3 is MiniMax's flagship model. Through late July and early August 2026 it spread across creative platforms — available on Leonardo, in Luma Agents, and via Pika's API Club.

Who makes MiniMax H3?

MiniMax H3 is developed by MiniMax. This page tracks its official announcements as they happen.

Where is MiniMax H3 available?

Based on official announcements tracked here, it is available via: Luma AI, Pika, Leonardo.

What is the latest MiniMax H3 news?

MiniMax H3 now available in Luma Agents (August 6, 2026). MiniMax H3 video generation is now integrated into Luma Agents, enabling up to 15 seconds of 2K video with native stereo sound, guided by text, image, video, and audio references in a unified workflow.