Wan 3.0Wan 3.0: 30 seconds in one take, from one model
Wan 3.0 is Alibaba's All-in-One video model — text to video, image to video and reference to video folded into a single model, up to 30 seconds per take, with up to 10 reference images plus reference video and audio in the same request.
- Failed generations auto-refunded
- Credit packs never expire
- Every model's price is public
- Cancel anytime in one click
- 2–30 s
- Length per take
- up to 1080P
- Resolution
- 10 img · 5 vid · 5 aud
- Reference material
- 10 credits / s
- From
Official prompts, and what they produced
A courtyard of Shaolin novices drilling in formation, a 30-second one-take ride down a California boulevard, a plush spaceship outrunning a woolen whale, a cowgirl filling her horse up at a wasteland gas pump, and an apsara dancer in stone-green and vermilion silk before a towering golden Buddha deep in a grotto — every clip below is an official Alibaba sample and ships the complete prompt that made it. Hit Remix to load any prompt straight into the generator above and make it yours.
- Wan 3.0
- Text to video
- 15s
- 720P
At the stone-paved courtyard behind Shaolin Temple, a group of young bald Shaolin novice monks stand in a perfectly aligned formation. Upon the instructor’s command, all of them begin repeating kick training drills together: high kicks, side kicks, and spinning kicks, flowing continuously into each other with complete synchronization and precision. The camera cuts to a close-up of one young monk performing the Golden Rooster Stands on One Leg stance. He balances on one foot, hands pressed together in a prayer gesture, remaining completely still and steady for several minutes. The instructor walks over to him, carefully observing his posture. The tip of his staff gently taps the monk’s elbow to correct the position and refine the movement. The camera then cuts to a row of wooden training dummies along the wall. Several senior disciples are practicing against the wooden dummies, their arms moving swiftly and fluidly between the dummy arms, performing rapid defensive blocks and counterattacks. The entire training courtyard is filled only with the sounds of wooden impacts, powerful shouts, and echoes of martial practice reverberating through the temple grounds.
- Wan 3.0
- Text to video
- 30s
- 720P
A 30-second, 16:9 horizontal, cinematic high-intensity One-Take visual spectacle. The overall style is "West Coast street fantasy," a visual language that fuses the grain of 90s street skate video with the polish of a modern commercial blockbuster. Grade the whole piece in extreme cinematic "California sunshine": a high-saturation blue sky against hard shadows of palm trees under direct sun, with the heat of the asphalt, the metallic rattle of shopping-cart wheels, and a fearless teenage defiance hanging in the air. Perspective stretches continuously as the lead races forward; the camera builds momentum through extreme low-angle follows and high-dynamic physical fly-throughs. The lead is a street-cool teenager in a color-blocked striped shirt and a black baseball cap, whose expression flips from loose and lazy to breaking-the-limit exhilaration. In the second scene the lead is the same teenager, but he appears in the real dimension as an observer. Shot 1, 0-6s: Open on a close-up: the teenager lounges inside a red steel shopping cart, backed by California's signature straight boulevard and towering palm trees. As the music drops, the camera pulls back at high speed and sinks to ground level, hugging the wheel at an extreme low angle. The teenager starts plunging down the steep road, the cart bucking hard on the surface, cars flying backward on both sides, the camera catching an almost reckless sense of acceleration. Shot 2, 7-15s: No cut. The camera stays glued to the ground like a skater in pursuit; the instant the teenager clears a makeshift wooden ramp, it rides the momentum into a graceful parabolic lift. The cart goes airborne and the camera passes underneath it. A giant commercial billboard now floods the entire field of view, and the camera pushes toward its center with an improbable crosshair precision. Shot 3, 16-24s: No cut. The teenager and the cart punch straight into the giant billboard as if breaking the dimensional wall. On contact with the poster, the three-dimensional body flattens in a material transformation, with real paper-tear texture and colored glitch ink appearing in frame. The teenager holds his gliding pose but has become a flat piece of artwork on the billboard. The camera executes a 180-degree horizontal orbit here, then dives from height back down to the ground. Shot 4, 25-30s: No cut. The camera lands smoothly on the street directly beneath the billboard, where another "real" teenager appears. He stops, slowly pulls his cap brim down, and looks up at the frozen version of himself on the billboard with a faintly amused smile. The camera follows his eyeline into a fast zoom-in, finally settling on the torn hole beside the "WAN" lettering on the billboard. The score fuses high-energy hip-hop with the physical sound of film winding through a spool. The visual information is highly condensed and full of stylish rebellion.
- Wan 3.0
- Text to video
- 20s
- 720P
Generate a 20-second, 16:9 ultrawide aspect ratio plush space adventure clip. A spaceship made of plush fabric, buttons, zippers, and stuffing escapes at high speed through a universe composed of giant woolen planets, plush nebulae, and ragdoll-like creatures. Piloting the spaceship are three small plush animals: a rabbit, a fox, and a crow. A gigantic interstellar whale covered in deep blue fur, with eyes like two glass buttons, chases after the spaceship from the nebula. The entire piece utilizes ultra-fine plush materials, cinematic space scale, a contrast between adorable appearance and strong movement, exaggerated FPV camerawork, and realistic soft-body physics. 20-Second Dynamic Timeline 0-3 seconds | Zippered Spaceship Launch The camera is inside the plush spaceship; the cockpit is composed of stitching, buttons, and fabric instruments. Red warning buttons flash rapidly. The rabbit pulls down a giant zipper control lever, and the front hatch of the spaceship bursts open like a sack. The camera rapidly retreats from the cockpit, passing through a zippered hatch to the outside of the spaceship. The spaceship is launched from a knitted crater on a yarn-like planet. 3-6 seconds | High-speed orbit around the yarn-like planet The camera stays close to the tail of the spaceship, flying low along the planet's surface. The ground is woven from giant yarn: Knitted mountains sweep past below at high speed; loose ends of yarn sway like a forest; the spaceship dodges sharply to the left and right; the camera tilts and rolls continuously; the thrusters emit white, cotton-like contrails. 6-8.5 seconds | Giant woolly whale appears The woolly nebula suddenly splits to the sides, and a huge, deep blue woolly whale bursts out from behind. Long wool-like hairs on the whale's surface are blown backward by the high-speed airflow, and its glass-like button eyes reflect the planet's light. The camera quickly circles halfway around the spaceship, switching to an inverted perspective, magnifying the whale from behind the spaceship. 8.5–11 seconds | Crossing the Button Asteroid Belt The spacecraft enters an asteroid belt composed of thousands of giant buttons. The camera follows the spacecraft in a continuous high-speed S-shaped motion: Passing between two rotating buttons; Skipping along the edge of a button hole; Entering the interior of a transparent button; Emerging from the other side; The whale smashes the buttons with its mouth, creating an explosion of fabric scraps and threads. 11–13.5 seconds | Swallowed by the Whale Beast The path ahead is blocked, and the spacecraft suddenly turns towards the whale's open mouth. The inside of the whale's mouth is like a giant plush toy box: Soft yet enormous furry teeth; Floating blocks, plush toys, and bells collide continuously; The camera rolls continuously in the narrow space; The spacecraft sprints low, skimming the surface of the tongue; The fur on the tongue creates violent ripples. 13.5–16 seconds | Escape from the Back Zipper The spacecraft discovers a huge golden zipper inside the whale's body. The fox character presses a button, and the spaceship launches a hook to unzip the zipper. The camera speeds along the zipper teeth, opening it to reveal the outer starry sky. The spaceship bursts out from the whale's back, the camera simultaneously performing a backward roll. 16-18 seconds | Whale's Soft Body Collapses After the zipper opens, the whale slowly flattens like a plush toy whose stuffing has shrunk, but without injury. A large cloud of white cotton bursts from the zipper, forming a huge, soft shockwave. The camera and spaceship are propelled forward by the cotton wave. 18-20 seconds | Through the Nebula and a Freeze-in-Time Frame with the Hero The spaceship bursts out of the cotton cloud and enters a pinkish-purple fuzzy nebula. The camera rapidly circles the side of the spaceship to face forward, flying backward, revealing three plush characters excitedly embracing each other through the cockpit window. The spaceship flies overhead towards a colorful button-shaped sun, and the scene ends. Negative Restrictions No hard metal spaceships, no human characters, no realistic flesh-and-blood behemoths, no horrific plush whales, no fire explosions, use cotton wool and soft lighting for thrusters, and no plush materials disappearing during high-speed movement.
- Wan 3.0
- Text to video
- 30s
- 720P
1. Synopsis A "refueling" incident on a brightly colored future wasteland, full of deadpan humor and absurdity. At a forgotten, Route 66-style gas station, a bored attendant receives an unusual customer: a cool cowgirl riding a real horse. Just as the attendant writes her off as another passing lunatic, the cowgirl walks straight to the pump, picks up the nozzle, and fills her horse up with fresh green grass. The surreal sight drops the world-weary attendant into complete cognitive-dissonance paralysis. 2. Scene description and lighting aesthetic Visual style: photoreal, graded with high saturation and high brightness for a "clear-day" feel, so the absurd story plays out in a perfectly normal, even lovely environment. Color aesthetic: strictly follow a bright teal-and-orange scheme. Teal: a cloudless sky in bright, clear azure leaning teal. The main gas-station building is retro yellow-orange-blue and slightly sun-faded. The pump is faded vintage red. Orange: the ground is cracked orange parched earth. Distant mountains read as warm terracotta. The cowgirl's scarf or jacket is orange. Characters and setting: Cowgirl: sharp and cool, a woman of few words. Wide-brimmed cowboy hat. Hard-wearing work boots and jeans, carrying the dust of a long road. Horse: an ordinary-coated, entirely unbothered real horse, as if it has seen all of this before. Attendant: greasy blue coveralls, oversized black sunglasses, slouched in a chair out of boredom with a toothpick in his mouth. He carries every reaction shot in the film. Gas station: classic Route 66 style, with a small convenience store beside it and a faded cola vending machine at the door, thick with American retro. 3. Camera style The camera style is "calm observation and absurd close-up": Lean heavily on locked-off medium shots and slow lateral moves, recording the event from a very calm, objective point of view without steering the emotion. At each beat of absurdity, cut abruptly to an extreme close-up — the moment grass sprays from the nozzle, or the attendant's blank face — and let that jarring magnification carry the joke. 4. Shot breakdown 0s-8s: Prologue, a boring afternoon Extreme wide, locked off. Blue sky, blowing sand. A retro yellow-orange-blue gas station, lightly faded, stands alone by the highway. At its door an attendant in sunglasses dozes lazily in a chair. Everything is quiet enough to hear only the wind. Long shot. On the horizon, a cowgirl on horseback appears slowly and breaks the calm. Medium. Hooves wake the attendant; he pushes his sunglasses up, sits upright with irritation, and studies the odd pair. The toothpick in his mouth shifts. His face says "another one asking for directions." 8s-16s: Turn, straight-faced nonsense The cowgirl says nothing. She swings down cleanly, leads the horse, and walks straight to an old pump. The horse extends its head beside the pump like it has done this before. Attendant POV, faint fisheye: the camera mimics his view from behind the sunglasses, and the woman-and-horse pairing looks even stranger. He watches her actually pick up the nozzle. Close. His brow furrows; he decides she is having fun at his expense. He takes off the sunglasses, about to shout. Slow-motion close-up: at the exact instant the glasses come off, he sees her aim the nozzle at the horse's mouth and squeeze the trigger. 16s-24s: Break, the moment cognition collapses The ultimate reversal: Close-up: the nozzle. What sprays out is not gasoline but bursts of fresh, dewy green grass. Extreme close-up: cut hard to the attendant's dumbstruck face. His eyes are saucers, his mouth slightly open, the toothpick forgotten mid-air. His expression freezes as if his entire worldview collapsed in one second. Medium. The horse eats with relish, squints happily, then flicks its tail with satisfied force. The dust the tail kicks up lands squarely on the still-frozen attendant. 24s-30s: Resolution, gone in the dust "Filled up" with grass. The cowgirl hangs the nozzle back on the pump. She digs a few coins from her pocket, walks to the store door, and drops them clinking into a tin can on the counter. She glances back at the petrified attendant; under the hat brim, the corner of her mouth seems to lift slightly. Final freeze: she swings back into the saddle and rides off in a cloud of dust. The camera pushes slowly in on the attendant. He is still in that dumbstruck pose, face covered in dust, the sunglasses he just removed still pinched in his hand. The toothpick finally drops from his mouth with a small clack. Freeze on that face of a man questioning his life. The film ends in absurd silence, leaving only the sound of the toothpick rolling on the ground as a gust passes through.
- Wan 3.0
- Text to video
- 18s
- 720P
Single continuous shot, using a dynamic combination of follow-cam and orbiting perspectives, 16:9 aspect ratio, CG animated feature-film quality. The characters’ faces are refined, smooth, and polished, with no pores or skin texture. Facial features are clean, elegant, and beautifully simplified. Colors are rich and saturated, with soft, warm lighting and dramatic contrast. 8K ultra-HD, shallow depth of field, motion blur, with a warm gold and malachite-jade color palette, and warm golden beams of light inside a dark grotto. A Dunhuang-style musical score interweaves deep drumbeats and pipa strings, with the rhythm accelerating as the duel intensifies. Deep inside the dim grotto, the space is almost completely black. At the center stands a huge golden Buddha statue, glowing faintly with a warm amber light in the darkness, its compassionate face only vaguely visible. A crack in the dome allows a single warm golden shaft of light to pour down. Within the light beam, golden dust motes slowly swirl. The murals on the four walls are submerged in darkness, with only faint outlines of malachite green and cinnabar red still discernible. The grotto is vast and hollow, with a long, deep echo. Within the beam of light stands a Feitian dancer: she has an oval face, willow-leaf brows, and phoenix eyes with slightly upturned outer corners. Her skin is like warm polished jade, smooth and delicate. Her vermillion lips are slightly parted, and a small cinnabar floral mark adorns the center of her forehead. Her jet-black hair is arranged in a high coiled bun, with two loose strands hanging beside her temples. On her head is a three-tiered golden lotus crown inlaid with turquoise beads, with golden chain tassels hanging down to her ears. Three layered necklace strands drape over her chest: the top layer is a fine pearl chain, the middle layer alternates turquoise and agate beads, and the bottom layer features golden lotus pendants. She wears a malachite-green cross-collar short top with narrow sleeves, trimmed at the collar and cuffs with golden scrolling floral patterns. She wears amber-gold high-waisted wide-leg trousers, gathered at the ankles with golden rings. Around her waist is a wide cinnabar-red satin sash tied into a bow, with the ribbon ends hanging down to her knees. A pair of three-meter-long ribbons are attached to her shoulders—cinnabar red on the left, malachite green on the right. Each ribbon is about twenty centimeters wide, edged with a fine golden trim. When extended, they appear richly colored and elegantly fluid. She wears gold armlets engraved with lotus motifs on both upper arms, and three gold bangles on each wrist. She holds a red-lacquered pipa in reverse grip, painted with a golden phoenix, with scroll-cloud carvings on the headstock. She is barefoot, with golden bells tied around her ankles. Lurking in the darkness is a Vajra warrior: he has a broad, square face, large bulging eyes, a wide nose, and a downturned mouth exposing his lower fangs. Three vertical wrinkles mark his forehead, and his expression is fierce. His reddish-brown short hair rises upward like flames. His bare upper body is powerfully muscular, with deep bronze skin and ochre-gold highlights on the lit areas. Across his chest runs a wide ochre-gold brocade sash, slung diagonally from his left shoulder to his right ribs. On both arms are gold armbands engraved with furious dragon motifs, and on his forearms are dark brown bracers. Around his waist is a dark ochre battle skirt reaching above the knees, embossed with flame patterns and lined along the edges with bronze studs. His calves are wrapped with dark brown crisscross bindings. He is barefoot, with broad, heavy feet. 0–3 seconds, slow push-in + slight rise: The camera slowly pushes forward out of the darkness into the grotto, as the warm golden beam illuminates the center. Within the light, the Feitian begins dancing in the tribhanga pose, holding the pipa behind her back in reverse grip. Her cinnabar-red and malachite-green ribbons whirl with her spinning body, flying through the warm light and tracing smooth arcs, their gold edging glinting. The gold bells at her ankles jingle delicately. From the Buddha’s shadow, the Vajra warrior steps out with a heavy stride. His bare foot slams the ground with a thunderous impact that echoes and amplifies through the dome. The stone floor beneath him cracks outward in a radial pattern, while flakes of gold paint shake loose from the Buddha’s surface, spinning and falling through the light beam. The Feitian stops spinning and turns to face him. She brings the pipa from behind her back to the front, holding it horizontally. Her layered necklaces sway lightly with the motion, while the ribbons continue curling behind her with residual momentum. 3–7 seconds, arcing orbit + lateral tracking: The Vajra warrior lets out a low roar and charges forward, throwing a straight punch at the Feitian. The force of the punch pushes away the golden dust in the light beam, creating visible shockwave ripples that spread outward. The Feitian sidesteps in tribhanga form, gliding gracefully to evade. Her cinnabar-red ribbon lashes out, wrapping tightly around the Vajra warrior’s forearm, the gold trim flashing in the warm light. He yanks hard, pulling her closer. Using the force against him, the Feitian spins and raises the pipa horizontally to block his second punch. The soundboard cracks apart, and the strings snap with a sharp, trembling sound that echoes through the dome. Fragments of the golden phoenix painting scatter into the air. Using the rebound from the block, the Feitian leans back into an aerial flip. The hem of her malachite-green top and her flowing ribbons expand into brilliant arcs of color as she rotates. Both feet strike the Vajra warrior in the chest, snapping his ochre-gold brocade sash and sending it flying. The Vajra stumbles backward and crashes into the Buddha’s toes. More flakes of gold paint shake loose, spinning in the warm light like a rain of golden fragments. The Feitian lands steadily on one knee, still gripping the broken pipa, while her necklaces and gold bangles jingle sharply. 7–11 seconds, rapid push-pull + sudden rise to overhead: The Vajra warrior roars in rage and slams both fists into the ground. Shockwaves spread outward from the impact point, driving loose stones across the floor. The impact travels to the walls, causing the dark murals’ pigment to flake away in showers. The Feitian leaps onto the Buddha’s palm, using it as a springboard to launch herself into a spinning attack. Her cinnabar-red ribbon sweeps across the Vajra warrior’s face, its end striking his cheek. He grabs the ribbon and tears it apart with a violent tug. At the torn edge near the Feitian’s shoulder, frayed threads drift visibly in the warm light. The camera suddenly rises to the ceiling for a top-down view. Below, within the warm golden beam, the two figures clash—one soft and graceful, one hard and forceful—their rhythms interwoven. Ribbon fragments and gold paint flakes spin and float in the light, glittering as they drift. 11–15 seconds, high-speed dive + ground-level follow shot: The Feitian hurls the broken remains of the pipa toward the Vajra warrior’s face. He smashes it apart with a punch, sending red lacquer fragments and broken strings exploding through the warm light. Taking advantage of the brief visual obstruction, the Feitian spins in close and drives both palms into his waist and ribs. A heavy impact sound reverberates through the grotto. The Vajra’s massive body crashes into the base of the Buddha statue, causing one corner of the pedestal to collapse. The great Buddha slowly tilts, emitting a prolonged low groan of structural stress. Rising from the rubble, the Vajra launches one final charge. The Feitian flies forward with a frontal kick. At the very center of the light beam, the two strike each other simultaneously. The sound reaches its peak, then suddenly drops into a brief instant of silence. The collision sends debris flying everywhere. Within the warm golden shaft, gold paint fragments and torn ribbon pieces slowly spin and settle. Both fighters step back, stabilize themselves, and breathe heavily. The remaining ribbons continue twisting in the lingering aftershock, while bits of stone still fall with soft rattling sounds. Do not duplicate the same character in the frame. The Feitian dancer and the Vajra warrior must be completely different in body type and temperament. Their facial features must remain refined and undistorted, maintaining a smooth animated-character facial texture with no pores. Their body structures must be anatomically normal, and their costumes, jewelry, and styling must remain fully consistent throughout.
Two reference photos, one runway film
Reference mode is where Wan 3.0 earns its name. This official example feeds one photo of a model and one photo of a black embroidered qipao, then names them in the prompt — the gold peony-and-bird embroidery survives frame for frame in the finished take.
Prompt
The female model from Image1 wears the black embroidered sleeveless qipao from Image2 and walks on a high-end fashion runway show. The camera first rapidly scans across the golden thread details of the peony and bird embroidery on the qipao, capturing the delicate craftsmanship and luxurious texture. The shot then cuts to a full-body view of the model confidently walking from the end of the runway. Her steps are sharp, powerful, and elegant. The high slit of the qipao reveals the graceful lines of her legs with each step, while the silk satin fabric reflects a cold, refined sheen under the spotlight. The camera quickly pushes forward from the front into a half-body shot, showcasing the stand collar, traditional knot buttons, and the complete embroidery details across the chest. The model's expression is cold, confident, and proud. The camera circles around half a turn to capture her side silhouette and the dynamic movement of the high slit. Immediately after, a low-angle upward shot captures her walking toward the end of the runway and making a precise turn. The skirt hem slightly lifts with the rotation, creating an elegant motion. Finally, a frontal close-up captures her eyes directly looking into the camera. Camera flashes rapidly flicker, and the image freezes. Overall style: Fast-cut, sharp rhythm, premium fashion editorial aesthetic, luxury fashion film quality, cold-toned lighting, strong contrast atmosphere.
What is Wan 3.0?
Wan 3.0 is the video generation model from Alibaba's Tongyi Lab, in public beta since 6 August 2026. Its headline change is scope: where earlier Wan releases split text to video, image to video and reference generation across separate models, 3.0 folds them into one. The same model takes a prompt on its own, a first frame, or a pile of reference material, and the mode follows from what you hand it.
The second change is length. Wan 3.0 generates up to 30 seconds in a single pass, which is long enough for a continuous camera move or a four-beat shot list without stitching clips together. Sound is produced with the picture rather than added afterwards, and you decide per take whether you want it. Reference capacity is the third: up to 10 images, 5 videos and 5 audio clips in one request, referred to by number inside the prompt.
On Molyin, Wan 3.0 runs on the same credits as every other model in the video generator — no separate plan, no per-model subscription. You can compare it against everything else in the model library, and see what a take costs before you spend anything on the pricing page.
Wan 3.0 vs Seedance 2.5 vs MiniMax H3
Three long-form models, three different strengths. Wan 3.0 takes the widest reference intake and the cheapest volume tier; Seedance 2.5 matches its 30-second ceiling with more aspect ratios; MiniMax H3 trades length for a 2K master.
| Spec | Wan 3.0 | Seedance 2.5 | MiniMax H3 |
|---|---|---|---|
| Length per take | 2–30 s | 4–30 s | 4–15 s |
| Resolution tiers | 480P / 720P / 1080P | 480P / 720P / 1080P | 768P / 2K |
| Reference intake | 10 images + 5 videos + 5 audio | 9 images + 3 videos + 3 audio | 9 images + 3 videos + 3 audio |
| Audio | Optional, same price either way | Optional | Always on |
| Seed | Yes — repeat a take | No | No |
| Credits per second | 10 / 18 / 36 | 25 / 56 / 100 | 18 / 29 |
| Best for | Long single takes and reference-heavy work at volume pricing | Cinematic 30-second work with wide aspect ratios | Short takes that need a 2K master |
Rates shown are current execution prices on Molyin.
Wan 3.0 vs the models available on Molyin
| Model | What it is best at | Released | On Molyin |
|---|---|---|---|
| Wan 3.0 | 30-second takes, 10-image omni reference, 480P volume tier | Aug 6, 2026 (public beta) · open weights announced | Live — generate above |
| LTX 2.5 | Fast and Pro tiers up to 4K, camera-motion presets, 24–50 fps | Aug 11, 2026 · open weights | Live — generate above |
| Seedance 2.5 | Cinematic 30-second takes with six aspect ratios | Aug 7, 2026 (API) | Live — generate above |
| FLUX 3 | Video and image from one model, storyboard keyframes | Aug 7, 2026 (video API) | Live — generate above |
| Wan 2.7 | Instruction-based video editing, cheaper 1080P | Aug 3, 2026 | Live — generate above |
| MiniMax H3 | 2K master with sound always on | Jul 31, 2026 · open weights announced | Live — generate above |
| HappyHorse | Lip-sync dialogue and built-in video editing | Jun 2026 (1.1) | Live — generate above |
| Seedance 2.0 | Standard, Fast and Mini tiers, 4K on Standard | Apr 2026 (API) | Live — generate above |
| Kling 3.0 | Native dialogue in five languages, element references | Feb 4, 2026 | Live — generate above |
Release dates are the point each model became usable, not the announcement date.
Wan 3.0 full specifications
| Parameter | Value | Notes |
|---|---|---|
| Model | Wan 3.0 (Tongyi Wanxiang 3.0), Alibaba Cloud | All-in-One model — one model covers all three modes. |
| Modes | Text to video · Image to video · Reference to video | The mode follows from what you supply: a prompt alone, a first frame, or reference material. |
| Resolution | 480P / 720P / 1080P | 480P is the volume tier; 1080P is the finishing tier. |
| Length per take | 2–30 seconds | Continuous slider, not fixed steps. |
| Aspect ratios | auto · 16:9 · 9:16 · 1:1 · 4:3 · 3:4 | Auto follows your material and intent; the other five are explicit. |
| Audio | Optional, generated with the picture | Switching audio on or off does not change what a take costs. |
| Reference material | 10 images · 5 videos · 5 audio clips | Refer to them by number inside the prompt, in upload order. |
| Reference video budget | Reference seconds + output seconds ≤ 30 s | Longer reference video leaves less room for the take itself. |
| Document reference | Announced by Alibaba, not yet open on Molyin | Alibaba has announced doc, xls, ppt, pdf, md and web pages as reference material. It is not available here yet; this row will change when it is. |
| First and last frame | Supported in image to video | Give both ends and let the model interpolate between them. |
| Seed | Supported | Reuse a seed to come back to the same take. |
| Credits per second | 10–36 credits / s | Lowest is 480P, highest is 1080P. |
Everything listed here is already supported on the site, and stays in sync with the model's latest official capabilities.
What a Wan 3.0 take costs
Per-second credit pricing by resolution — the same credits that every other model on Molyin draws from.
| Model | Resolution | Credits/sec | 5s Video | 5s Video + 3s Reference |
|---|---|---|---|---|
| 480p | 10 | 50 credits | 80 credits | |
| 720p | 18 | 90 credits | 144 credits | |
| 1080p | 36 | 180 credits | 288 credits |
- Exception: reference video seconds for Wan 3.0 are billed at the full per-second rate.
Subscription plans and one-time credit packs are compared side by side on the pricing page.
How to generate with Wan 3.0
Three steps from idea to a finished take
Pick a model and a mode
Choose text to video, image to video with a first frame, or reference mode with up to 10 images plus reference video and audio.
Write the prompt
For anything past 10 seconds, break the take into timed beats — 0-6s, 7-15s, 16-24s — the way the official samples on this page do, and name each reference by number where it should appear.
Generate and export
Pick 480P, 720P or 1080P and any length from 2 to 30 seconds, decide whether you want sound, preview in place, then download the original file.
Writing a 30-second take
Thirty seconds is long enough that a one-line prompt runs out of instructions halfway through. The Wan 3.0 guide breaks it into a timed shot list — here is the formula it opens with.
[Global Setup] + [Start-End Seconds] + [Picture Content & Camera] + optional [Transition]
Time Slice
Explicit start-end seconds (e.g. 0-5s, 5-12s). Cutting 30 seconds into 3-5 segments is the most stable.
Picture & Camera
What happens in this slice and how the camera moves — action load matched to the slice length.
Global Setup
An opening paragraph that locks style, color tone, mood and hard no-gos for the whole clip.
Transition
How slices join: natural cut, dissolve, wipe — just name it.
Wan 3.0 questions
Capabilities, pricing and how it differs from Wan 2.7
Start Creating with Molyin Today
Sign up free and turn your first idea into a cinematic video in minutes.