Choose the example that matches your shot: latte art and a glass pavilion use MiniMax H3 at 2K; bouquet wrapping and a paper boat use H3 Max at 768P. Each includes the video, full prompt, original request settings and actual output dimensions. For a portrait coffee-shop clip, generate with the latte prompt → and select 5 seconds, 2K and 9:16.
| Your shot | Model | Output framing | Original request settings |
|---|---|---|---|
| Coffee-shop latte art | H3 | Portrait, 1440×2560 | 5 seconds, 2K, 9:16 |
| Florist wrapping a bouquet | H3 Max | Landscape, 1344×768; requested 9:16 | 6 seconds, 768P, 9:16 |
| Architectural pavilion reveal | H3 | Landscape, 2560×1440 | 5 seconds, 2K, 16:9 |
| Paper-boat atmosphere shot | H3 Max | Landscape, 1344×768 | 5 seconds, 768P, 16:9 |
Latte Art
Request: 5 seconds, 2K, 9:16. Output: 1440×2560, about 5.2 seconds.
A barista's hand pours steamed milk from a stainless steel pitcher into a black espresso cup held at a slight angle on a light grey marble countertop. The milk stream settles and blooms into a full rosetta (fern-leaf) latte-art pattern by the end of the pour. Warm, directional side light rakes across the cup and pitcher. Vertical 9:16 framing, macro-close perspective, shallow depth of field, no camera movement, no text, no logos, no additional hands or objects entering frame.
Generate with this prompt → Select 5 seconds, 2K and 9:16.
Make it yours: swap the cup, countertop and latte-art pattern; keep a clear endpoint—the finished pattern—and the fixed close-up camera.
Bouquet Wrapping
Request: 6 seconds, 768P, 9:16. Output: 1344×768 landscape, about 6.6 seconds. H3 Max text-only generation currently produces landscape output; use a frame image with the desired ratio or crop or pad the clip for portrait delivery.
A florist's hands wrap a bouquet of ranunculus and eucalyptus in brown kraft
paper on a sunlit marble counter, then tie it with twine in a neat bow; soft
morning window light, shallow depth of field, close-up vertical framing,
warm cozy mood, gentle paper-rustling motion
Generate with this prompt → Select 6 seconds and 768P; the original 9:16 request produced the landscape clip above.
Make it yours: replace the flowers and wrapping material, keeping the wrap-then-tie sequence and the focus on the hands and bouquet.
Glass Pavilion
Request: 5 seconds, 2K, 16:9. Output: 2560×1440, about 5.2 seconds.
Read this longer prompt in five parts: setting, sculpture motion, camera path, materials and lighting, then sound. Copy the full text before adapting those parts.
Single continuous cinematic shot inside a vast glass pavilion suspended above a stormy ocean at blue hour. A kinetic sculpture made of hundreds of translucent crystal ribbons rises from a mirror-black reflecting pool and unfurls in a slow spiral, refracting warm amber light through cool teal mist. The camera begins inches above rippling water, glides forward, then cranes upward into a wide orbital reveal of the pavilion and distant lightning. Physically accurate reflections and refractions, volumetric directional lighting, crisp microdroplets, controlled natural motion, stable geometry, seamless temporal consistency, premium luxury commercial finish, high dynamic range, subtle film grain, no people, no logos, no text, no cuts, no flicker. Ambient sound: low ocean rumble, soft rain, delicate glass resonance, no speech.
Generate with this prompt → Select 5 seconds, 2K and 16:9.
Make it yours: replace the building and central installation; retain the water-level opening, forward glide and rising reveal, with ambient sound described at the end.
Paper Boat
Request: 5 seconds, 768P, 16:9. Output: 1344×768 landscape, about 5.2 seconds.
After a morning rain, a red paper boat drifts through shallow water beneath a stone bridge. A low camera tracks smoothly beside it, small ripples spread across the water, and birds sing softly in the distance. No text or subtitles.
Generate with this prompt → Select 5 seconds, 768P and 16:9.
Make it yours: swap the boat color and bridge setting; retain the low, smooth tracking shot and let ripples and distant birds provide the background.
Questions Before You Adapt a Prompt
Which reference inputs can I use?
Use first_frame_image or last_frame_image for frame control. H3 requires a first frame when using a last frame and also accepts image_urls, video_urls and audio_urls reference lists; audio references need an image or video, and frame inputs are separate from the reference lists.
H3 Max supports a first frame or last frame independently; omit aspect_ratio when supplying either, and use images with the ratio you want. Max does not accept those reference lists. See the H3 parameter docs and H3 Max parameter docs.
How should I make a portrait H3 Max clip?
Text-only generation currently produces landscape output: the bouquet request specified 9:16 but returned 1344×768. The documented portrait path uses a first- or last-frame image with your desired ratio and omits aspect_ratio; alternatively, crop or pad the landscape result in an editor. The H3 latte example above already has a portrait output at 1440×2560.
What does the generate link prefill?
It fills in the full prompt only. Select duration, resolution and aspect ratio yourself, and supply any images separately; each example lists its original request settings beside the video.
Use the H3 and H3 Max pricing guide to budget your batch, or the H3 Max docs for API fields. Start with one subject, one action and one camera path: generate the paper-boat shot → and select 5 seconds, 768P and 16:9.
