Gallery
Real MiniMax Hailuo H3 product video samples.
11 clips from the H3 tiers we run, each with the exact prompt that produced it. H3 is the model behind our photo-to-product-video tool: audio is generated with the video, and an image-to-video pass keeps the product, logo and colour from your photo.
These are model reference samples. Register to prepare and save editing prompts in your workspace. Video rendering is not available yet.
Clips
Every tier, one sample each.
Press play to stream. Duration and resolution are measured from the file. Portrait and landscape samples sit in the same 16:9 frame.
MiniMax H3-Developer Text-to-Video
Prompt A young woman stands alone on a quiet beach during a stormy sunset, holding an old photograph in her hand. Waves crash against the shore as strong wind moves her hair and coat. She slowly looks toward the horizon, then releases the photograph into the wind. The camera begins behind her shoulder and slowly circles around to reveal her emotional expression. Dark clouds gradually open as warm sunlight breaks through. Natural ocean movement, realistic wind simulation, soft backlight, shallow depth of field, muted cinematic color palette, subtle film grain, intimate emotional storytelling, poetic f
MiniMax H3-Developer Image-to-Video
Prompt The race car speeds forward.
MiniMax H3-Developer Reference-to-Video
Prompt The race car driver raises a hand in celebration of the victory.
MiniMax H3 Max Turbo Text-to-Video
Prompt A lone hooded traveler stands motionless in the middle of an endless pale desert, facing a gigantic futuristic spacecraft resting silently on the sand dunes. The enormous spacecraft begins to slowly hover above the ground, its massive shadow gradually separating from the sand as fine dust and grains of sand are lifted into the air beneath it. The traveler’s long robe gently moves in the desert wind. The camera slowly pushes forward from behind the traveler, emphasizing the extreme scale contrast between the tiny human figure and the colossal spacecraft. Subtle atmospheric haze moves across the
MiniMax H3 Max Turbo Image-to-Video
Prompt A lone mountaineer slowly climbs along a narrow snow-covered ridge high above the clouds, carefully planting his trekking poles into the deep snow with every step. Strong alpine winds blow through the scene, snow particles and fine powder swirling around his body, his hood and clothing moving naturally in the wind. The climber pauses briefly, catches his breath, then continues steadily toward the summit. In the background, a gigantic snow-covered mountain peak gradually emerges from thick clouds and drifting mist. The clouds move slowly across the mountain, occasionally revealing the sharp sum
MiniMax H3 Text-to-Video
Prompt Cinematic medium close-up of a desert warrior wearing a high-tech dust mask and stillsuit, glowing piercing bright blue eyes (eyes of Ibad). The wind blows fine dust across their face. Warm desert sunlight casting sharp side shadows. Cinematic focus, shallow depth of field, slow-motion film grain, photorealistic, Dune style, 8k.
MiniMax H3 Image-to-Video
Prompt The warrior raised the sword in his hand high and brandished it.
MiniMax H3 Reference-to-Video
Prompt Little Bear straightened his clothes, getting ready to look cool.
MiniMax H3 Fast Text-to-Video
Prompt A lone man walks through a rain-soaked neon city at midnight, reflections shimmering across the wet streets, distant headlights cutting through the mist, his coat gently swaying in the wind. The camera slowly tracks backward in front of him, subtle handheld movement, shallow depth of field, cinematic lighting, atmospheric fog, anamorphic lens flares, realistic rain, moody blue and amber color grading, dramatic cinematic composition, film grain, high-end feature film cinematography, 24fps.
MiniMax H3 Fast Image-to-Video
Prompt A tiny fluffy pale-blue creature sits alone in a minimalist white and light-blue space. Its soft fur gently sways in a subtle breeze. The creature slowly blinks its large round eyes and curiously looks around, then gently wiggles its tiny feet while its little tail sways lazily from side to side. After a moment, it looks up toward the camera and gives a soft, innocent smile. The camera slowly pushes in from a low-angle close-up, emphasizing its fluffy fur, round face, and delicate facial expressions. Soft natural light falls from the left, creating a subtle glowing shadow on the floor, with a
MiniMax H3 Fast Reference-to-Video
Prompt A futuristic silver Formula-style concept race car rests on an endless mirror-like salt flat beneath a dreamy pastel sunset sky. Massive pink and lavender clouds slowly drift across the horizon, perfectly reflected on the glassy surface below. The camera begins with an extremely low-angle wide shot, slowly dollying toward the car, then smoothly transitions into a cinematic orbit around the vehicle, revealing its sculpted aerodynamic body, polished chrome surfaces, exposed wheels, and flawless reflection. Subtle wind creates delicate ripples across the reflective ground while clouds move slowly
Tiers
Which H3 should you run?
- MiniMax H3-Developer — $0.030/sec
- MiniMax H3 Max Turbo — $0.036/sec
- MiniMax H3 — $0.057/sec
- MiniMax H3 Fast — $0.066/sec
Iterate on the cheapest tier that produces the motion you want, then re-render the final cut on a higher tier. Full per-second pricing for all 328 models we run is on the model catalog.
FAQ
Short answers.
Are these clips actually generated by MiniMax Hailuo H3?
Yes. Each clip is a real sample output from the tier named under it, served from this site. The prompt that produced it is shown verbatim below the clip, so you can judge how closely the output follows the brief.
What is the difference between the H3 tiers?
Same model family, different cost and speed trade-offs. H3 Developer is the cheapest per second and is where you iterate; H3 Fast is speed-optimised; H3 is the balanced tier; H3 Max and H3 Max Turbo are the highest-quality routes. All of them generate audio with the video rather than needing a separate scoring pass.
Can I turn a product photo into a video with H3?
Yes — that is the main use here. Start from a clean product photo (a white-background cutout or a relit lifestyle shot), use image-to-video, and describe one camera move: a slow push-in, a turntable orbit or a light sweep across the label. The photo locks the product, logo and colour; the prompt only adds motion.
Why does the same prompt come out differently?
Generation is stochastic. Lock what matters — product, background, ratio, duration and the one motion — and change one variable per pass. Rewriting the whole prompt between runs is what makes camera work and packaging details drift.
How do I generate my own?
Register or sign in to edit product photos with Qwen Image Edit. Other models on this page are catalog references.