MINIMAX H3 / THE SAMPLE GALLERY

See what’s possible.
Then make it yours.

Start free. See what PRO adds.

Explore four original fantasy scenes generated on JollyAI. Watch each result, inspect its prompt and try your own idea. PRO, image-to-video and reference-to-video comparisons are still being prepared.

Original MiniMax H3 cinematic scene of a desert train approaching a portalYOUR IDEA. IN MOTION.Watch the cinematic samples ↗
4 published videos3 creative workflows4 published storiesMore samples in preparation

FREE / START EXPLORING

A small idea. A real result.

Four-second presets for quick concepts and first experiments. Daily allowance and queue availability apply.

Text and image workflows have limited free access. Reference-to-video currently requires PRO.

Create a free video ↗

PRO / GIVE THE SHOT ROOM

More time. More detail.

Current public PRO duration options include 6, 8 and 10 seconds. Check the studio for available presets. PRO comparison samples are being prepared.

Reference-guided creation requires PRO. Check the studio for current presets and account access.

Explore PRO plans ↗

SAME IDEAS / DIFFERENT POSSIBILITIES

Watch. Compare. Create.

Click to play with sound.
Nothing autoplays.

T2V

Text to video

Start with an idea. Direct the subject, action, camera and sound.

Open workflow ↗
FREE4-second preset

The last train to another world

864 × 480

Generation prompt

Epic science-fiction cinema, photoreal live-action VFX. One sleek dark passenger train with a narrow acid-green light stripe races along straight rails across a vast red desert toward a gigantic circular turquoise light portal spanning the track. Huge sandstone cliffs, golden-hour sun, long shadows and a billowing trail of orange dust. A low aerial tracking camera follows the train from behind and beside it, keeping its locomotive and the portal legible. The train surges into the ring; light rolls across the carriages and reveals a star-filled blue world through the opening. Continuous coherent forward travel, solid rail geometry, one train and one portal. Sound: rhythmic steel wheels, deep electric engine, rising ethereal portal hum and a final airy rush. Tell one compact visual moment: the action starts immediately, followed by a clear magical reveal. One continuous cinematic shot. No lettering, logos, subtitles or watermarks. Landscape 16:9.

Generation settings
Workflow
Text to video
Seed
20260920
Render time
2.1 min for this example
Generate with T2V ↗
FREE4-second preset

An ocean above the clouds

864 × 480

Generation prompt

Dreamlike fantasy cinema with realistic volumetric light. One enormous translucent blue whale floats gracefully through a sea of towering white clouds at sunrise; a tiny unoccupied brass airship hangs nearby to reveal the whale's immense scale. Sunbeams pass through its glasslike body, revealing delicate luminous veins. The whale sweeps its tail and rolls gently beside the airship, sending a soft wave of cloud vapor around it. A sweeping wide camera tracks with the whale and reveals a rainbow in the cloud mist. Calm majestic physical motion, consistent whale anatomy, one whale. Sound: deep resonant whale song, rushing wind and delicate glass harmonics. Tell one compact visual moment: the action starts immediately, followed by a clear magical reveal. One continuous cinematic shot. No lettering, logos, subtitles or watermarks. Landscape 16:9.

Generation settings
Workflow
Text to video
Seed
20260921
Render time
1.6 min for this example
Generate with T2V ↗
FREE4-second preset

The garden that woke up

864 × 480

Generation prompt

Cinematic science-fiction discovery, photoreal film lighting. A single adult astronaut in a weathered ivory spacesuit with an opaque gold visor stands inside a vast dark cavern filled with dormant alien plants. A tiny blue bioluminescent seed floats just above one gloved palm. As the astronaut slowly opens the hand, the seed lifts into the air and a luminous turquoise wave spreads across the plants, transforming the cavern into a glowing garden. The camera pushes past the astronaut's shoulder to reveal the illuminated landscape. One astronaut, consistent suit and helmet, clear hand silhouette, believable movement. Sound: quiet breathing inside a helmet, a soft rising chime and the rustle of awakening plants. Tell one compact visual moment: the action starts immediately, followed by a clear magical reveal. One continuous cinematic shot. No lettering, logos, subtitles or watermarks. Landscape 16:9.

Generation settings
Workflow
Text to video
Seed
20260922
Render time
2.3 min for this example
Generate with T2V ↗
FREE4-second preset

One spark for the sleeping city

864 × 480

Generation prompt

Original cinematic stop-motion-style fantasy with detailed miniature craftsmanship. A small copper clockwork fox with a teal scarf and a glowing amber heart trots onto an empty rain-soaked cobblestone street in a dark miniature city. The fox taps a brass streetlamp with one front paw; a warm light turns on, followed by a chain of glowing windows and streetlamps receding into the city. A low camera follows the fox then gently rises to reveal the wave of light reflected in puddles. The fox looks up, ears lifting with wonder. One fox, consistent four-legged anatomy and copper mechanical design. Sound: tiny ticking gears, gentle rain, a bright electrical chime and soft city ambience. Tell one compact visual moment: the action starts immediately, followed by a clear magical reveal. One continuous cinematic shot. No lettering, logos, subtitles or watermarks. Landscape 16:9.

Generation settings
Workflow
Text to video
Seed
20260923
Render time
2.4 min for this example
Generate with T2V ↗
I2V

Image to video

Start with a still. See how H3 brings the same source image to life.

Open workflow ↗

Examples for this workflow are being prepared. No completed sample is published here yet.

R2V

Reference to video

Guide a new shot with a reference image for appearance and identity.

Open workflow ↗

Examples for this workflow are being prepared. No completed sample is published here yet.

WHAT YOU’RE COMPARING

Actual outputs. Visible settings.

These are curated, AI-generated JollyAI examples, not a guarantee that every prompt will produce the same quality. This initial collection contains four Free text-to-video examples. Matched PRO, image and reference comparisons will be added after generation and review.

Current PRO presets offer longer durations. Videos are encoded for web playback without artificial frame extension or resolution upscaling. Small duration differences come from the model’s frame alignment. Render times describe these runs and exclude queue waiting; they are not delivery promises.

All examples use JollyAI’s locally hosted H3 workflow. MiniMax’s separate 2K API workflow is not used here. Read the model’s official capabilities ↗

YOUR NEXT SHOT STARTS HERE

Try the idea.
Make the upgrade yours.

Start with a free text or image video. Choose PRO when your idea needs a longer shot.

GENERATION / ANIMATION PROMPT

Full prompt