Model · MiniMax · Hailuo AI & H3 · Via API · Independent
MiniMax H3 is the newest video model from MiniMax, the company behind the Hailuo AI app: text to video, image to video and reference to video with up to 9 images and 3 audio clips, at up to 2K. It has a public API, but it isn’t open on PictureToVideo yet — join the waitlist and we’ll email you the day it is. What runs here today is Hailuo 02, the physics-focused Hailuo AI model: 6- or 10-second clips at 768p for 4 credits per second, or 6-second 1080p Pro clips for 7, from a prompt or a photo. PictureToVideo is independent and not affiliated with MiniMax.
H3 lengths here aren’t set yet; the estimate uses the same 6 and 10 seconds as Hailuo 02.
Coming soon here
Reference-driven action with native stereo sound, up to 2K
≈ 36 cr
6s · 768p · 6 cr/s · ≈ $0.49 on Pro
A per-second price for H3 here isn’t confirmed yet. The 6 credits per second shown is a placeholder until it opens here, and only then do we publish the real rate.
Live here now · same clip, closest settings
MiniMax · live now
6s · 768p · silent
24 cr
12 less than H3
Hailuo 02 Standard: physics-heavy action from a prompt or a photo, 6 or 10 seconds, silent.
Same 1080p frame
6s · 1080p · silent
42 cr
6 more than H3
Hailuo 02 Pro renders 1080p in 6-second clips only, so a 10-second request falls back to 6.
Keep a character
6s · 1080p · sound
114 cr
78 more than H3
Reference to video is H3’s headline feature. Until it opens, start Kling 3.0 from a photo of your character or product: it keeps that look through the clip, with sound when audio is on.
H3 isn’t open on PictureToVideo yet. Leave your email and we’ll send one message the day it goes live here, with its confirmed price per second.
No spam · one email when MiniMax H3 opens here
1366×768 · 6 or 10s
24 cr
1920×1080 · 6s
42 cr
Standard · 768p for this shot.
On a phone feed, 768p is plenty and leaves 39 more clips in a Pro month than 1080p would.
Verdict
Sports and acrobatics: the physics-heavy motion it was built around. A 6-second 768p draft costs 24 credits.
“A parkour athlete in a grey hoodie sprints across a rooftop, vaults a low wall and rolls out of the landing, side tracking shot, late afternoon light”
Suggestions, not test scores: fit notes follow MiniMax’s positioning of Hailuo 02 and its limits here (silent, 16:9, 6 or 10 seconds).
Switch the generator to Image, upload a photo you have the rights to and describe what moves: the dog leaps, the wave breaks, the scarf whips in the wind. The Hailuo AI model starts from your photo and renders a silent 16:9 clip.
Hailuo AI renders 16:9 here. A wide photo keeps everything in frame; a portrait photo gets reframed, so crop it to landscape first — or use Kling 3.0 for vertical video.
Hailuo AI image to video costs the same per second as text to video: 4 credits at 768p, 7 at 1080p. A 6-second 768p clip is 24 credits (≈ $0.33 on Pro).
MiniMax H3’s reference to video takes up to 9 images and 3 audio clips in one job, for characters, products and voices that should stay consistent. Until it opens here, start Kling 3.0 from a photo of your character or product to keep its look through the clip.
Model facts: MiniMax’s Hailuo 02 announcement and our API provider’s Hailuo 02 Standard and Pro model pages, checked October 14, 2026; MiniMax H3 from public H3 API listings, checked October 5, 2026. Credits: our live price list.
| Hailuo 02 | Wan 2.5 | Kling 2.5 Turbo | Kling 3.0 | |
|---|---|---|---|---|
| Maker | MiniMax | Alibaba | Kuaishou | Kuaishou |
| Clip length | 6 or 10s | 5 or 10s | 5 or 10s | 3–15s |
| Resolution | 768p · 1080p | 720p · 1080p | 720p · 1080p | 720p · 1080p · 4K |
| Aspect ratios | 16:9 | 16:9 · 9:16 · 1:1 | 16:9 · 9:16 · 1:1 | 16:9 · 9:16 · 1:1 |
| Native audio | No | No | No | Optional |
| Lowest cr/s | 4 | 9 | 6 | 10 |
| One short clip | 24 cr (6s 768p) | 45 cr (5s 720p) | 30 cr (5s 720p) | 60 cr (6s 720p) |
| Reach for it when | Action and physics | Square, 720p drafts | Cheapest drafts | Long takes, audio, 4K |
A woman in a brown jacket walks a clifftop path above the sea at golden hour, waves break white over the rocks below, the grass sways in the wind, static wide shot
A blank sheet of paper on a wooden table lifts, creases and folds itself into an origami swan, crisp paper edges, soft window light, static high angle
A golden retriever bounds through a ripe wheat field at sunset, stalks whip past, dust glows in the backlight, ears flapping, slow motion, low tracking shot
A tabby cat sprints upright on its hind legs down a red running track in a sunny stadium, front paws pumping, tail swinging for balance, side tracking shot
An amber perfume bottle with a gold cap stands on a black pedestal, spotlight beams cut through soft haze, the glass glints as the light shifts, slow push-in
A motorcyclist in black leathers races a sport bike along an elevated city highway at night, lightning splits the clouds, neon signs streak past, chase shot from behind
Say what drives the motion: a gust of wind, a hard landing, a wave breaking, a push. Hailuo AI reads physics better when the cause is in the prompt, not just the result.
Describe the whole beat — sprints, vaults the wall, rolls out of the landing — so the clip finishes the move inside 6 or 10 seconds instead of stopping mid-air.
One move per clip: a tracking shot, a slow push-in or a static wide. Let the subject do the work; a busy camera on top of fast action is harder to read.
Find the take at 768p for 24 credits, then re-render it as a 6-second 1080p Pro clip for 42. The prompt optimizer runs on both, so the wording carries over.
| Model | Resolution | Credits / sec | With audio | Shortest clip | Longest clip |
|---|---|---|---|---|---|
| Hailuo 02 | 768p | 4 | — | 6s · 24 cr | 10s · 40 cr ≈ $0.54 |
| Hailuo 02 | 1080p | 7 | — | 6s · 42 cr | 6s · 42 cr ≈ $0.57 |
| MiniMax H3Coming soon · placeholder | 768p | 6 | Included | 6s · 36 cr | 10s · 60 cr ≈ $0.82 |
| MiniMax H3Coming soon · placeholder | 1080p | 9 | Included | 6s · 54 cr | 10s · 90 cr ≈ $1.22 |
| Wan 2.5 | 720p | 9 | — | 5s · 45 cr | 10s · 90 cr ≈ $1.22 |
| Wan 2.5 | 1080p | 14 | — | 5s · 70 cr | 10s · 140 cr ≈ $1.90 |
| Kling 2.5 Turbo | 720p | 6 | — | 5s · 30 cr | 10s · 60 cr ≈ $0.82 |
| Kling 2.5 Turbo | 1080p | 6 | — | 5s · 30 cr | 10s · 60 cr ≈ $0.82 |
Example: three 6-second 768p Hailuo AI drafts (72 credits) plus one 6-second 1080p Pro final (42 credits) is 114 credits, ≈ $1.55 on the Pro plan. Pro rows show 6 seconds as both shortest and longest — it is the only length the 1080p tier renders. The MiniMax H3 row is a placeholder rate until H3 opens here.
Still curious? Email support@picturetovideo.org — a human replies within a day.
Draft action on Hailuo AI today, and join the MiniMax H3 waitlist for the day it opens. Starter pack from $2.99, no subscription.
