Bodies that move correctly
MiniMax built Hailuo around physics: gymnastics, dance and fast footwork keep their weight and timing. Hailuo 2.3 made complex body movement more fluid again.
MiniMax Hailuo · by MiniMax
Hailuo is MiniMax's family of video models, known for human motion that holds together, from dance and gymnastics to the small movements of a face. Its newest generation, H3, adds sound and readable text in the same pass. Write a scene below and it opens in Video Studio on MiniMax H3 Max.
Example video
Opens in Video Studio with the model you picked. You see the price in credits before anything runs.
MiniMax built Hailuo around physics: gymnastics, dance and fast footwork keep their weight and timing. Hailuo 2.3 made complex body movement more fluid again.
Hailuo 2.3 was tuned for natural micro-expressions in live action: a frown that softens, eyes that well up. H3 carries that into clips with sound.
H3 generates voice, effects and music with the picture, and MiniMax says it renders text and brand names accurately. Our chalkboard sign came out spelled right.
Hailuo 2.3 improved anime, illustration and game-CG looks, so stylised characters move with the same physics as live action.
Every clip below was generated with the prompt shown, unedited. Load any prompt into the generator and make your own version.
Vertical selfie-style video: a beauty creator in her twenties with a sleek black bob sits at a vanity with ring-light reflections in her eyes. She holds up a lipstick to the camera, swipes it on, presses her lips together and says with a grin, "okay, this one's the winner." Natural skin texture, warm bedroom light. Sound: her voice, a soft click of the lipstick cap, quiet room tone, no music.
Vertical shot in an empty concrete parking garage at night: a street dancer in a white t-shirt and baggy cargo pants hits a sharp hip-hop routine, popping his chest, sliding into a quick footwork sequence and freezing on the beat with one hand on the floor. The camera stays low and steady at full body. Fluorescent lights hum overhead. Sound: a punchy hip-hop beat, sneakers squeaking on concrete.
Vertical close-up: a man in his fifties with a grey beard reads a text message on his phone in a dim kitchen. His expression shifts slowly from worry to disbelief to quiet relief; his eyes glisten and he lets out a long breath, pressing a hand to his mouth. Very subtle, natural micro-expressions. Soft light from the phone screen. Sound: a refrigerator hum, his breath.
Vertical fashion film: a tall model with cropped silver hair walks straight toward the camera down a minimalist white runway, wearing a flowing emerald silk gown that ripples with every step. Photographers' flashes pop at the sides. The camera slowly pulls back to keep her full body in frame. Sound: camera shutters, heels on the runway, a low electronic beat.
Vertical shot outside a small corner coffee shop in the morning: a smiling barista in a green apron sets down a chalkboard sign on the sidewalk and turns it toward the camera. The sign reads "OPEN LATE · FRESH CROISSANTS" in neat white chalk letters. She wipes her hands on the apron and waves to a passer-by. Sound: street ambience, a bicycle bell, her cheerful "morning!".
A gymnast in a navy leotard performs a back handspring on a balance beam in a bright arena, lands cleanly, steadies her arms and holds the pose while the crowd applauds. Smooth tracking camera at beam height, realistic body weight and timing.
Anime style, hand-drawn look: a schoolgirl with a red scarf stands on a rooftop at sunset as the wind lifts her hair and scarf. She turns to the camera with a determined smile while cherry blossom petals swirl past. Soft cel shading, warm orange sky, gentle camera push-in.
Start here
Most clips
The newest generation, tuned for quality: sound, text and up to 15 seconds at 1080p. Start here.
5 s clip 145 credits · 5–15 s
Open with MiniMax H3 MaxOption 2
Drafts and volume
Half the credits and faster, tuned for prompt adherence. Find the shot here, then re-run it on H3 Max.
5 s clip 73 credits · 5–15 s
Open with MiniMax H3 Max TurboOption 3
Cheap silent clips and anime
The previous generation at 768p: strong body motion and stylised looks, no sound.
From 51 credits · 6–10 s
Open with MiniMax Hailuo 2.3 StandardHailuo 2.3 Pro, Hailuo 02 and MiniMax H3 Reference to Video are also in the studio's model menu. Credits are read from the live catalog; the Standard plan ($24/month) includes 2,000 credits, about 13 five-second MiniMax H3 Max clips.
Write the shot in the box at the top of this page, or start from an example. Add a photo to use it as the first frame.
Your prompt, photo, format and the Hailuo version you picked carry over. Nothing runs until you press Generate.
5, 10 or 15 seconds on H3. The credit price shows under the Generate button before anything is spent.
Find the take on H3 Max Turbo, then run the same prompt on H3 Max.
Human motion: dance, sport and the small movements of a face. The newest generation, H3, also generates sound and readable text with the picture.
H3 is MiniMax's third-generation video model, sometimes called Hailuo 3, released in July 2026. H3 Max is a quality-tuned version of it made by fal, with Turbo as its faster, cheaper tier.
H3 Max and H3 Max Turbo do: voice, effects and music come with the picture. Hailuo 2.3 and 02 clips are mostly silent.
Up to 15 seconds on H3, and 6 or 10 seconds on Hailuo 2.3. Join several clips in the Video Editor for a longer scene.
No. Every run costs credits, priced by clip length and resolution and shown before you generate. Credits come with a monthly plan and unused credits roll over up to the plan's cap.
Yes. Paid Affogato plans include a commercial license and no watermark, within our terms of service.
Facts checked November 15, 2026. Sources: MiniMax: H3 · MiniMax: Hailuo 2.3 · MiniMax: Hailuo 02 · fal: H3 Max · fal: MiniMax H3 explained · MiniMax: video generation guide · Kapwing: how to prompt MiniMax H3