MiniMax H3
Sharp 2K clips from 5 to 15 seconds, made from text, one or two frames, or reference images and video.
- Resolution
- 768P / 2K
- Length
- 5–15 seconds
- Sound
- Generated with the video
- Cost
- 100 credits for 5 seconds at 2K
Describe a shot or drop in a photo. EveryGen AI writes the prompt, picks the model and renders a clip with sound, then you refine it in the same chat.
Pick a model in the composer, or let the EveryGen AI agent choose one for the shot. Both render sound together with the picture.
Sharp 2K clips from 5 to 15 seconds, made from text, one or two frames, or reference images and video.
The longest takes in EveryGen AI: 4 to 30 seconds at up to 720p, from text, frames or reference images, video and audio.
Type what should happen on screen. EveryGen AI expands it into a full shot description, with subject, camera, light and pace, before it renders.
Upload a photo or still as the opening frame, add a closing frame if you have one, and EveryGen AI animates the move between them.
Attach images, clips or audio as guidance for a character, style, motion or rhythm, and EveryGen AI keeps the new shot on brief.
Start from one of these, or describe your own idea in any language.
Performance and mood shots that match the energy of your track.
“A singer under a single spotlight on a dark stage, slow orbiting camera, haze and soft lens flares.”
Product hero shots, unboxings and UGC-style clips for paid social.
“A perfume bottle on wet black stone, water droplets, slow macro push-in, luxury commercial lighting.”
Vertical 9:16 clips built to land in the first second of the feed.
“Vertical 9:16: a cat jumps onto a kitchen counter and knocks a cup off in slow motion.”
Anime, claymation, 3D and cartoon looks from a single sentence.
“Anime style: a girl on a rooftop at sunset, wind in her hair, city lights switching on below.”
Plan a sequence in chat, then render it one shot at a time.
“Plan a three-shot opening for a sci-fi short film, then render the first shot.”
Bring portraits, old photos and product stills to life.
“Animate this portrait with a subtle smile and a slow camera push-in.”
Browse clips by style, open one to watch it, then use its prompt as your starting point.
Type an idea in any language, or attach images, clips or audio.
EveryGen AI turns your words into a detailed shot prompt and picks the right tool and model.
The clip renders with audio in a few minutes. Close the page or press stop and it keeps rendering, then lands in your chat and library.
Ask for a slower camera, a new angle or a 9:16 version. Every take stays in the conversation.
Every clip comes with generated sound and effects, so there is no separate audio step.
Change one thing at a time in plain words instead of rewriting a long prompt.
Earlier versions stay in the chat and your library, so you can always go back.
Leave the page or press stop: started renders keep going and are delivered when they are ready.
Plans come with monthly credits. What a clip costs depends on the model, length and resolution, and the composer shows it before you send.
$19.90/month
2,000 credits a month
About 20 five-second MiniMax H3 clips at 2K
$49.90/month
5,500 credits a month
About 55 five-second MiniMax H3 clips at 2K
$99.90/month
11,500 credits a month
About 115 five-second MiniMax H3 clips at 2K
$199.90/month
24,000 credits a month
About 240 five-second MiniMax H3 clips at 2K
Yearly plans charge 10 months for 12 months of credits. Credit packs work without a plan.
EveryGen AI (every-genai.com) is an AI video generator you use through a chat. You describe a shot or upload an image, and its video agent writes the prompt, renders the clip with sound on models such as Seedance 2.5 and MiniMax H3, and lets you refine it in the same conversation.
Right now MiniMax H3 (up to 2K, 5 to 15 seconds) and Seedance 2.5 (up to 720p, 4 to 30 seconds). Both render sound with the video, and new models are added to the composer as they become available.
Each render costs credits based on the model, length and resolution, and the composer shows the cost before you send. For example, a 5-second MiniMax H3 clip at 2K costs 100 credits. If a render fails, its credits are refunded automatically.
Yes. Upload a photo as the opening frame, optionally add a closing frame, and describe the motion. You can also use images as references for a character or style instead of as frames.
Yes. MiniMax H3 and Seedance 2.5 both generate audio together with the picture.
The render keeps going. When it finishes, the clip appears in the chat and in your library.
If you were looking for the chat-based AI video generator at every-genai.com, you are in the right place. EveryGen AI, EveryGenAI and Every Gen AI are all names for this site. We are not the Everygen.ai service formerly known as Viewmax.
No. EveryGen AI at every-genai.com is an independent product. It is not affiliated with Everygen.ai or Viewmax LLC, and it is not an official site of MiniMax, ByteDance or any other model provider.
Last updated November 10, 2026
Describe the first shot and watch it render with sound.