A Chinese company called MiniMax just released H3, an AI tool that creates and edits video from a simple text request. If you have ever wanted to make a short clip without a camera or editing skills, MiniMax H3 is the kind of tool that puts that within reach.
The Gist
- MiniMax H3 is a new AI model that generates and edits video, released on July 31.
- It scores as the top tool for AI video editing, but trails rivals on turning text into video.
- You can try it through MiniMax’s own app, Hailuo AI, without downloading anything.
Have ChatGPT Recap This Article
ChatGPTMeet MiniMax H3, a fresh AI video maker
An AI video model is software that builds moving images for you, either from a written description or from a photo you feed it. MiniMax H3 is the newest one, and it comes from a Chinese company that has been quietly building tools like this for a while.
MiniMax says H3 was built for real commercial work, the kind used in film, advertising, video games and product design. That is a step up from a toy that makes funny five-second clips, and it signals the tool is aiming at people who actually publish video.
The company also stressed it cared about hardware compatibility, meaning the model is meant to run without needing the most expensive setup. For a beginner, that is the difference between a tool you can only read about and one you can actually open and use.
What “open weights” means for you
H3 is being released as an open-weights model, and the full version lands in the coming days. Open weights means the company shares the actual model so others can download and run it, instead of locking it behind a private service.
There is a catch worth knowing. It is free for personal use, and businesses can use it too as long as they earn under twenty million dollars a year and give credit. If that sounds abstract, we break the whole idea down in our plain guide to what an open-weight model is.
For you, the practical effect is choice. Open tools tend to spread fast, show up inside other apps, and push prices down, which is good news whether or not you ever touch the raw model yourself.
Keep learning on AI Noobies:
- Google Earth Just Pulled Its Fake Image Tool
- Which Free AI Chatbot Wins in 2026 for Beginners
- AI Slop on LinkedIn Now Has a Report Button
Where H3 wins and where it trails
Independent testing gives a mixed but honest picture, and that is useful to see. H3 ranks as the strongest tool for video editing, meaning changing and polishing footage you already have.
It is weaker at the flashier trick. When it comes to turning plain text into a brand-new video, it trails Google’s Gemini Omni Flash, and it sits behind both that model and Seedance 2.0 for turning a still photo into a moving clip. If you want to explore the Google side, we covered Gemini Omni and its video features too.
The takeaway is not that one tool wins everything. Different models are good at different jobs right now, so the smart move is picking the one that fits the task in front of you.
How to try AI video like H3 this week
You do not need to download the model or understand any of the technical parts. MiniMax runs a consumer app called Hailuo AI where you can type a scene and watch it build a short clip.
Start small and specific. Instead of asking for something vague, describe one clear shot, like a calm beach at sunset with gentle waves, and see how close the result lands. A short, precise request almost always beats a long, fuzzy one.
The bigger picture is that AI video is getting cheaper and easier every month. Trying one clip now, even a clumsy one, is the fastest way to understand what these tools can and cannot do for you.
Stay tuned on AI Noobies.
About the Author
