// ENGINE 08 — MINIMAX
Hailuo 2.3
Cinematic precision with emotional depth. Emotional performance — expressive faces and fluid human motion, framed by the model itself.
- CREDITS
- 40 cr
- QUALITY
- good
- SPEED
- fast
- DURATIONS
- 6 / 10s
- ASPECTS
- auto
- IMAGE INPUT
- first frame
GENERATE
— renders with hailuo 2.3 — 40 crPROMPT SAMPLES
— written for this engine — use one as a starting pointPROMPT_01
An old fisherman reads a letter on the dock at dawn; his weathered face moves from confusion to quiet joy as the light rises.
PROMPT_02
A ballet dancer rehearses alone in a dim studio, breath visible, movements building from tentative to full-out.
PROMPT_03
Close-up of a young woman hearing her song on the radio for the first time — disbelief, then laughter, eyes brimming.
PROMPT_04
A street violinist plays in falling snow, fingers moving fast, passersby slowing one by one to listen.
PROMPT_05
A father teaches his daughter to ride a bike; the moment he lets go, pride and terror cross his face in the same second.
PROMPT_06
A boxer between rounds: exhaustion, then her coach says something off-screen and resolve rebuilds her posture piece by piece.
ABOUT
Hailuo 2.3 is MiniMax's video engine, tuned for the human element: micro-expressions, gesture, and the kind of fluid body motion that makes dance and performance footage feel alive. Its shots read as felt rather than staged, which is rarer than it sounds in generated video.
The specialty is the transition — a face changing its mind mid-shot. Grief to relief, doubt to a smile, disbelief to laughter: Hailuo renders the in-between frames of an emotion with a subtlety the spec-sheet engines don't attempt. Bodies get the same treatment; choreography stays liquid through fast, complex movement.
It's also the studio's most opinionated framer — there is no aspect-ratio setting, because the model composes each 6 or 10-second shot itself. Describe the emotion and the moment; Hailuo decides where the camera should be. At 40 credits with fast renders, it's an affordable way to get performances no amount of camera direction produces elsewhere. When the line has to be spoken out loud, that's Kling 3 territory.
STRENGTHS
— what this engine does better than the others- Micro-expression: faces that change their mind convincingly mid-shot
- Fluid, liquid body motion through fast and complex choreography
- Emotional arcs rendered in one take — the in-between frames of a feeling
- Self-composed framing: the model does its own cinematography
- Fast renders at a mid-tier 40-credit price
PROMPTING TIPS
— how to talk to this model- TIP_01Write the emotional arc, not the shot: "confusion to quiet joy" outperforms any lens direction here.
- TIP_02Skip camera language entirely — Hailuo frames the shot itself, and fights prompts that micromanage it.
- TIP_03Anchor the feeling in a physical event (a letter, a song on the radio) so the expression has something to react to.
- TIP_04For dance, name the build — "tentative to full-out" — and the choreography will carry the progression.
USE CASES
01
Performance and dance
Choreography, gesture, physical expression — motion stays fluid through fast, complex movement.
02
Emotional close-ups
A face changing its mind mid-shot: grief to relief, doubt to a smile. Micro-expression is the specialty.
03
Music-driven moments
Mood pieces and lyric-video beats where the feel matters more than the framing spec.
04
Character studies
Short portraits of a person in their element — let the model's own composition instincts do the cinematography.