Minimax H3 is a general-purpose, multimodal video generation model available through ElevenLabs. It supports text-to-video generation, image animation, reference-guided creation, multi-shot generation, and video editing with natural-language prompts.

What can Minimax H3 generate?

Minimax H3 generates video from text prompts, animates images, connects opening and closing frames, and uses subject, style, or motion references to guide the result.

How does Minimax H3 edit existing videos?

Minimax H3 edits existing videos using natural-language instructions. Upload your video to ElevenLabs, then describe changes such as replacing an object, redirecting movement, or adjusting the camera.

How do I use Minimax H3 on ElevenLabs?

Select Minimax H3 from the ElevenLabs model library, enter a prompt, and add optional image or video references. Generate the clip, then revise your prompt to refine the result.

Is Minimax H3 free on ElevenLabs?

No. Minimax H3 video generation on ElevenLabs requires a paid plan; there is no free video generation.

Can I add a voiceover to Minimax H3 videos?

Yes. Minimax H3 generates the visuals, and you can add an ElevenLabs voiceover afterward in your preferred language.

Generate beautiful assets with Minimax H3

Generate Video

Trusted by 1M+ users

Minimax H3 unifies video generation, reference control, and natural-language editing in one model.

Minimax H3

A playful but premium 15-second motion graphic. Champagne split-flap tiles ripple from the center, flip into geometric MOTION, then lock together as burgundy and black machined letterforms under a gallery spotlight.

A masked shrine runner, white owl mask, silver braid, navy coat traced in gold kintsugi, lantern in hand. Burning paper talismans, then a sprint through flooded torii gates as the village burns. Night, firelight on water, cinematic anime, ink-wash title cards. SCROLL OF THE BURNING SHRINE.

General-purpose multimodal video generation

Generate Video

Create videos from text, images, or reference clips, then edit objects, actions, and camera direction with natural language.

One model for multiple video tasks

Create controlled video from text or images, then edit the result with the same model.

Generalized reference control

Guide subjects, style, or motion with reference media while maintaining a coherent generated scene.

Edit videos with natural-language prompts

Describe changes in plain language to replace objects, redirect action, or adjust the camera.

Native multi-shot storytelling

Create connected shots with consistent subjects, settings, and visual direction throughout the sequence.

In-context regeneration

Regenerate selected moments so new actions fit the surrounding clip.

Faithful prompt following and visual detail

Follow detailed prompts while preserving fine visual details and rendering readable text and brand elements.

Create with Minimax H3 with full control

Select Minimax H3

Choose MiniMax H3 from the ElevenLabs video model library to start a multimodal generation or editing project.

Enter prompt and add references

Describe the scene, then optionally add images or video clips to guide the subject, style, motion, or transition.

Generate and refine

Generate the clip, review it, then refine the action, objects, or camera direction with clear instructions.

Write a prompt...

Minimax H3

Minimax H3 Max

Gemini Omni Flash 1.1

Recraft V4 Styles

720p

16:9

Frequently asked questions

Get started with Minimax H3

The best image, video, and audio models — all inside ElevenCreative. Start generating today.

Sign up

Discover more image & video generation models

Explore our full library of AI image and video generation models, each with unique strengths and capabilities.

Minimax H3 Max
Minimax H3 Max generates polished, prompt-faithful video faster than real time on ElevenLabs.
Gemini Omni Flash 1.1
Gemini Omni Flash 1.1 combines fast video generation and conversational editing with precise cinematic control.
Recraft V4 Styles
Recraft V4 Styles turns visual references into style-consistent raster and vector images.
Flux 3 Video
Flux 3 Video generates grounded, multimodal AI video from prompts, images, and clips.
Seedance 2.5
Seedance 2.5 generates coherent 30-second videos with reference control and targeted edits.
Seedream 5.0 Pro
Seedream 5.0 Pro creates controllable commercial images, dense layouts, and realistic product visuals.
Wan 2.5 Video
Wan 2.5 Video generates short, high-resolution videos from text or images with smooth motion and flexible control.
Seedream 5 Lite
Seedream 5 Lite rapidly generates images up to 3K with deep reasoning, precise prompt-following
HeyGen Avatar 4
HeyGen Avatar 4 quickly generates expressive, lifelike talking avatars from a single photo
Creatify Aurora
Creatify Aurora transforms a single photo and audio into ultra-realistic AI avatar videos with nuanced motion and visual style
Veed Fabric 1.0
Veed Fabric 1.0 creates realistic talking videos from any image, with lifelike gestures, lip sync, and flexible style control.
Sync 3
Sync 3 offers 4K lip synced video generation with shot-level realism, emotion preservation
Gemini Omni Flash
Gemini Omni Flash edits and creates short videos through step-by-step natural language direction.
GPT Image 2
Seedance 2.0
Google Nano Banana 2
Runway Gen-4.5
LTX 2 Pro
Wan 2.5
Seedream 4.5
Seedream 4
Seedance 1.5 Pro
Seedance 1 Pro
OmniHuman 1.5
Topaz Upscale
FLUX.2 [Pro]
FLUX.1 Kontext [Pro]
Kling 3.0
Kling O3 Pro
Kling 2.6
OpenAI Sora 2 Pro
Kling 2.5
GPT Image 1.5
OpenAI Sora 2
Google Nano Banana Pro
Google Nano Banana
Google Veo 3.1 Fast
Google Veo 3 Fast
Google Veo 3