Model
.ai
tracking open ai models

Minimax H3: t2va (T2V) – Polyhedron: Perfect Eyes // Perfect Skin // Perfect Hands

8569°
Minimax H3: t2va (T2V) – Polyhedron: Perfect Eyes // Perfect Skin // Perfect Hands
Created by
Polyhedron_AI
on Civitai
by Polyhedron_AI - Minimax H3: t2va (T2V) – Polyhedron: Perfect Eyes // Perfect Skin // Perfect Hands
Creator notes
TL;DR: This LoRA helps with skin texture, the eyes and the hands for Minimax H3: t2va. It's the final LoRA model and might be already overtrained ... so use it with lower weights. Just experiment! Okay, folks,  I don't really know what to say other than: MiniMax H3 is ridiculously good. <3 The hype is real. I've spent the last few days messing around with both models:   ref2va (reference/image to video) [Get ready to meet your maker!] t2va (text to video) [here]   So here's my first LoRA — the bread-and-butter one: skin, hands and eyes. And I hit a wall almost immediately. We need to be a bit careful with LoRAs here, because it's the same story as with Wan 2.2: the base model already knows(!) what all of this is supposed to look like, and that can clash. Anyway. They do have an effect, but the whole thing is very prompt-sensitive. Keep that in mind.   The current videos were made with four LoRAs, two of mine, two external:   *polyhedron_minimax_h3_image_lora_v2__nippl8es // (currently only up on my Patreon — we'll be needing this one later, wink wink)   *polyhedron_minimax_h3_image_lora_v2__perfe8ct // (this is the LoRA here in its final version; if it feels overtrained to you, all the training steps are on my Patreon as well with the workflows I used for the example videos)   And then two external ones. First:     That's what got me started on this LoRA in the first place — though the workflow is an absolute mess. So I'm putting up a clean, tidied-up one that by now runs entirely on Polyhedron Suite nodes. The quality speaks for itself. Go try it. I've also gone and pulled the whole MiniMax H3 universe into my own repo, which you can grab here — updated today — and it's available through the ComfyUI Manager too:     And here's the second LoRA:     Tiddies! Already in there, and we'll be needing that one later as well. :D   We've been hearing "this model is the real gamechanger" for years now. This time I actually think it's true. :)   Catch you later!   Frank
Trigger words
perfe8ct
perfect hands
perfect skin
perfect eyes
2,975 downloads · 126 likes on Civitai
SHA256
95A3950EFD58AF267652EACA52CF76F5E94FB93FCF72CC7B42EBFA3A1FF34666
Ranking badge
Updated Oct 22, 2026
Minimax H3: t2va (T2V) – Polyhedron: Perfect Eyes // Perfect Skin // Perfect Hands on ModelTrends.ai: #146 Hottest AI model
Made this model? Show this badge on your page.
Paste it into your Hugging Face card or Civitai description.
Also ranked: #6 Hottest MiniMax H3 model · #6 Hottest MiniMax H3 LoRA
Not this creator?
Minimax H3: t2va (T2V) – Polyhedron: Perfect Eyes // Perfect Skin // Perfect Hands Popularity Trend
Heat score since 29 Aug 2026
Heat score
8,569°
This week
▲
3.5%
8,561° then, 8,569° now
This month
▲
23.3%
8,523° then, 8,569° now
Past year
n/a
Wasn't released/tracked
Minimax H3: t2va (T2V) – Polyhedron: Perfect Ey…: Heat score since 29 Aug 20267,000°7,439°7,879°8,318°8,757°30 Aug9 Sep20 Sep30 Sep12 Oct22 Oct8,569°ModelTrends.ai
7,000°
Floor
The scale runs from 0° to 9,999°. Everything below 7,000° is kept for models no source reports on yet, which carry a stand-in number there so lists stay complete. 7,000° is where measured data begins: a model with real data but almost no downloads sits at the floor, and the measurements fill the degrees above it. The floor says nothing about the model; only the degrees above it do, and they are what separates one model from another.
+1,061°
Civitai downloads
2,975 downloads of this exact version on Civitai. Downloads are the heaviest ingredient and are counted on a log scale: ten times the downloads adds the same 300°, so a model with a million downloads is not a thousand times hotter than one with a thousand. The six download sources share that one term, split between them in proportion to their counts.
+316°
Likes
126 likes and thumbs up across the sources. A like weighs half a download, on the same log scale.
+92°
Adoption rate
About 16 downloads a day, measured between two readings of the sources, never guessed from a lifetime average. Speed weighs a quarter of the download term, so a fresh surge shows without letting a one-day spike outrank years of use.
+100°
Family boost
+100° because this is a LoRA of a family this site is nudging forward right now. It is a bounded nudge inside the score and is listed here openly; it can never hide a model or outweigh a thousandfold difference in downloads.
Show detailed breakdown
Every number on this site can be checked. This model's heat score is 8,569°, and here it is built one ingredient at a time, with the running total after each: the floor, the degrees each source adds, the bonus for the kind of model and what the age of its weights takes away. The colors are the bands of the chart above. The glossary at the end defines every term.
  1. 1
    Start at the floor.
    Why 7,000° and not zero: the scale runs from 0° to 9,999°, and the part below 7,000° is reserved for models no source reports on yet, which carry a stand-in number there so the lists stay complete and never outrank a model with data. 7,000° is where measured data begins. A model with real data but almost no downloads sits here; everything that follows is measurement, and the room above the floor holds ten tenfold steps in downloads at 300° each. The floor is the same for every model, so it decides nothing; the degrees above it do.
    7,000°
    running total
  2. 2
    Add the downloads.
    2,975 download-equivalents in all, from Civitai downloads. Ten times the downloads adds the same 300°, so size matters but never runs away with the score.
    +1,061°
    8,061°
    running total
  3. 3
    Add the likes.
    126 likes and thumbs up, at half the weight of a download.
    +316°
    8,377°
    running total
  4. 4
    Add the adoption rate.
    About 16 downloads a day right now, measured between two readings, at a quarter of the weight.
    +92°
    8,469°
    running total
  5. 5
    Family boost.
    An open, bounded nudge for LoRAs of a family this site is pushing forward right now.
    +100°
    8,569°
    running total
  6. =
    The heat score.
    Where this model stands on every list of this site. Nothing else adjusts it.
    8,569°
Glossary: the terms used here
Heat score
One number for how much a model is used and wanted right now, from 0° to 9,999°. Every list on this site that says hottest is sorted by it and nothing else.
Floor
The 7,000° every model with real data starts from. The scale runs from 0° to 9,999°; the part below the floor is reserved for stand-ins (models without data), so that a model with data always ranks above one without. The floor is the same for everyone and decides nothing; the degrees above it are the measurement.
Download-equivalents
Everything the score treats as a download, added up: downloads themselves, each torrent seeder as 1,000, each AI Horde image and each run on a generation site as one.
Log scale
Counting by tens, not by ones: going from 1,000 to 10,000 downloads adds the same degrees as going from 10,000 to 100,000. It keeps giants from crowding out everything else.
Likes
Thumbs up and likes from every source, worth half a download each on the log scale.
Adoption rate
How many downloads a day a model gets right now, from the difference between two readings of a source. A quarter of the weight of the download term.
Measured and estimated
Measured means two readings of a source exist and the difference is real. Estimated means only today's total is known and the past is spread over the model's life; the site says which one it is showing.
Seeder
Someone who keeps a copy of the weights online for others to download over a torrent, or a hosted mirror doing the same. Each one counts as 1,000 downloads.
AI Horde
A volunteer network that generates images for free on donated computers. The number of images it has made with a model is public, and each counts as one download.
Generation sites
Sites that host their own copies of a model and let people generate with it, such as PixAI, SeaArt and Yodayo. Their downloads count like Civitai's and each generation run as one download.
OpenRouter tokens
A text model served through an API, such as Claude or GPT, has no weights to download, so its usage is measured in tokens: the pieces of text it read and wrote through OpenRouter, a gateway many apps use, over the last 30 days. Every 10,000,000 tokens count as one download. Source: OpenRouter (openrouter.ai/rankings), CC BY 4.0.
Full model (checkpoint)
The weights themselves, the file that generates on its own. It gets a bonus because add-ons depend on it.
LoRA and embedding
Small add-on files that steer someone else's weights toward a style, a character or a concept. They cannot run alone.
Family and weights
The base a model is built on, such as SD 1.5, SDXL, Flux or Wan. Every model inherits the release date of its family's weights.
Age
Degrees taken off for how long ago the family's weights came out: 150° a year, up to 750°. It keeps old downloads that had years to pile up from freezing the top of the list.
Family boost
A small, bounded nudge for LoRAs of a family this site is pushing forward right now, always listed in the breakdown.
Stand-in score
A stable number below the floor given to a model no source reports on yet, so lists stay complete. It never ranks above a model with data.
New! (grace period)
A model less than 7 days old shows New! instead of a score and holds no place in the rankings. Its score is computed from day one, but a model one day old grows faster than anything, so without the wait the week's #1 would always be whatever came out yesterday rather than what held up for a week.
Trend
How a model's growth over a period compares with the typical model's growth over the same period. Faster than typical reads as a rise, slower as a fall.
Typical model
The middle of the catalog: half the models with a figure grew faster over the period, half slower.
How this model's popularity has moved against the rest of the catalog: its growth in downloads, likes and usage over each period, measured against the typical model's growth over the same period, with its heat score at the time. Above the typical growth reads as a rise, below it as a fall. The past is reconstructed from today's totals, the model's first day and its measured download speed where two readings exist; a model that did not exist at the start of a period shows n/a.
Example renders
Prompts and settings shared by the people who made these renders on Civitai. Pick a render to see how it was made.
Render by Polyhedron_AI on
Civitai
·
Prompt on BunchaPrompts
Flag
Settings
Size
1882x1074
Prompt
// SUBJECT a 24 year old woman sits by the window, an ordinary everyday face, not a model, // FACE slightly irregular features, one eyebrow set a little higher than the other, one upper eyelid heavier than the other, a small bump on the bridge of her nose, a thin upper lip over a fuller lower lip, a narrow jaw, a smooth rounded chin, no cleft chin, a face that is not quite balanced, // EYES green eyes looking towards the viewer, round iris, depth in iris, small dark pupils, catchlights in eyes, faint blood vessels in sclera, natural eye moisture,
Settings
Size
1882x1074
Prompt
// SUBJECT a 24 year old woman sits by the window, an ordinary everyday face, not a model, // FACE slightly irregular features, one eyebrow set a little higher than the other, one upper eyelid heavier than the other, a small bump on the bridge of her nose, a thin upper lip over a fuller lower lip, a narrow jaw, a smooth rounded chin, no cleft chin, a face that is not quite balanced, // EYES green eyes looking towards the viewer, round iris, depth in iris, small dark pupils, catchlights in eyes, faint blood vessels in sclera, natural eye moisture,
Settings
Size
1882x1074
Prompt
// SUBJECT a 24 year old woman sits by the window, an ordinary everyday face, not a model, // FACE slightly irregular features, one eyebrow set a little higher than the other, one upper eyelid heavier than the other, a small bump on the bridge of her nose, a thin upper lip over a fuller lower lip, a narrow jaw, a smooth rounded chin, no cleft chin, a face that is not quite balanced, // EYES green eyes looking towards the viewer, round iris, depth in iris, small dark pupils, catchlights in eyes, faint blood vessels in sclera, natural eye moisture,
Render by Lionprincess on
Civitai
·
Prompt on BunchaPrompts
Flag
Settings
Sampler
Euler
Steps
8
Guidance
1
Size
1216x2176
Prompt
Generate a 15-second 16:9 photorealistic fantasy video. Image 1: use as the exact character appearance, hairstyle, crown, purple-and-gold clothing, jewelry, glowing triangular symbols, and dark temple environment reference. Preserve her recognizable face, golden hair, costume design, and magical color palette while translating the illustrated reference into realistic cinematic fantasy. [0s–2.5s] Medium frontal shot inside the shadowy ancient temple. The sorceress kneels exactly as in Image 1 with eyes closed and both palms raised. Her long golden hair moves gently in an unseen magical wind. Realistic warm firelight outlines her hair while violet energy coils slowly around both hands. Camera makes a controlled slow push toward her face. The glowing golden triangular sigil above her head flickers to life. [2.5s–5s] Extreme close-up on her face. Her eyelids open, revealing softly glowing violet irises. Tiny reflections of the golden sigil appear in her eyes. She speaks calmly but with absolute authority: Sorceress: "The seal has broken." As she finishes the line, a deep vibration travels through the temple and fine dust falls from the stone arch behind her. [5s–7.5s] Hard cut to a low-angle medium-wide shot. She rises smoothly to her feet while keeping both hands extended. The two violet energy streams whip outward, curve behind her, and form a rotating circular arc of runes. Her golden hair lifts dramatically from the magical pressure. The camera slowly circles clockwise around her as stone tiles begin glowing with matching triangular patterns. [7.5s–10s] Close tracking shot from her right side. She turns toward the dark doorway behind her as a violent pulse of purple light erupts from the corridor. Her expression changes from serenity to focused determination. She closes one hand into a fist, causing the floating violet rune-circle to collapse into a concentrated sphere of energy above her palm. [10s–12.5s] Wide cinematic shot. A massive shadowy presence begins forming deep inside the doorway, visible only as towering movement and two distant glowing eyes. The sorceress steps forward and thrusts her glowing hand toward it. She declares: Sorceress: "Then let the old gods remember my name." The violet sphere launches forward as a narrow beam, illuminating the entire temple in purple and gold. [12.5s–15s] The beam strikes the darkness with a thunderous flash. Camera rapidly pulls backward as golden and violet shockwaves race across the floor and up the columns. For the final second, the light clears enough to reveal the sorceress standing unwavering at the center of the temple, hair flowing, both magical sigils blazing brighter than before. End on her intense expression as the music hits a final powerful orchestral chord. Audio: Epic cinematic fantasy score beginning with low cellos, distant choir, and deep ceremonial drums. At 2.5s, reduce the music slightly beneath the dialogue: Sorceress: "The seal has broken." Add a deep stone rumble immediately afterward. From 5s, introduce heavier percussion, rising brass, and shimmering magical textures. At 7.5s, add a sharp supernatural pulse from the dark doorway, followed by a concentrated magical hum as the energy sphere forms. At 10s, briefly pull the music back beneath the second line: Sorceress: "Then let the old gods remember my name." At 12.5s, synchronize the beam impact with a powerful magical blast, stone reverberation, rushing energy, and full orchestra with choir. Finish at 15s with one massive brass-and-choir chord and a fading low-frequency rumble. Environmental ambience throughout: cavernous temple reverb, faint wind, falling dust, subtle crackling magical energy. Continuity: Keep the sorceress identical to Image 1 throughout: same adult face, golden wavy hairstyle, jeweled crown, purple-and-gold ceremonial outfit, white gloves, earrings, belt ornament, and triangular magical symbolism. Preserve the same dark stone temple architecture, lighting direction, violet magical energy, and gold illumination across every shot. The violet energy generated from her hands must remain visually consistent in color and texture, and the golden triangular sigil above her must remain the same symbol throughout.
Settings
Size
1882x1074
Prompt
// SUBJECT a 24 year old woman sits by the window, an ordinary everyday face, not a model, // FACE slightly irregular features, one eyebrow set a little higher than the other, one upper eyelid heavier than the other, a small bump on the bridge of her nose, a thin upper lip over a fuller lower lip, a narrow jaw, a smooth rounded chin, no cleft chin, a face that is not quite balanced, // EYES green eyes looking towards the viewer, round iris, depth in iris, small dark pupils, catchlights in eyes, faint blood vessels in sclera, natural eye moisture,
Render by Lionprincess on
Civitai
·
Prompt on BunchaPrompts
Flag
Settings
Sampler
Euler
Steps
8
Guidance
1
Size
1120x832
Prompt
For the target video, at 0.00 seconds into the target video, <Picture 1> (from [Shot 1]) is fully referenced.JGLWLK integrated_multimodal_description: [Shot 1] Live-action, cinematic, natural soft lighting, a close-up portrait of the young woman from <Picture 1> fills the frame, preserving her wavy blonde hair, light freckles, blue-green eyes, glossy lips, silver hoop earrings, layered silver chain necklaces with heart and ring pendants. The camera slowly pulls out with medium amplitude at slow speed, revealing She is wearing a double slit dress. burgundy in color that drapes smoothly over her figure, the fabric catching warm natural light from large penthouse windows. She stands in a trendy Pinterest-inspired penthouse living area with clean modern furniture, soft neutral textiles, trailing plants, and floor-to-ceiling glass. With a warm, inviting smile she turns slightly, the silky dress shifting fluidly against her body as she runs one hand lightly down the side of the fabric in a natural, sensual motion. The sweet-voiced young woman (S1) says with gentle enthusiasm: <d>[English] Hey everyone… come closer, I have to show you this dress.</d> She takes a soft step forward, the camera tracking with her at slow speed while she lifts the hem slightly to let the light glide across the glossy burgundy silk, then lets it fall again, her posture relaxed and open. [Shot 2] At 00:06.500, the camera cuts to a medium shot as she turns fully toward the lens, the dress flowing with the movement. She places both hands lightly on her waist, fingertips tracing the smooth fabric, and continues with a soft, inviting tone: <d>[English] It’s so soft… feels like liquid when you move. Perfect for a quiet evening in.</d> She gives a small, genuine smile, tilting her head slightly as natural daylight wraps around her. overall_soundscape: Soft room tone of a quiet modern apartment with faint distant city ambience filtering through open windows. Gentle rustle of silky fabric as she moves, light footsteps on polished floor, and the quiet shift of jewelry chains. non_diegetic_music: Soft ambient electronic pads at a slow tempo with warm, sparse piano notes that gently rise and fade under her speech.
Render by Lionprincess on
Civitai
·
Prompt on BunchaPrompts
Flag
Settings
Steps
8
Guidance
1
Size
608x1088
Prompt
integrated_multimodal_description: [Shot 1] 2D-animated, cinematic anime style, a dynamic low-angle medium-wide shot fully matches <Picture 1>: a young woman with voluminous curly blonde hair held by a blue plaid headband, blue eyes, fair skin, wearing a fitted blue sailor-collar dress with white trim and a dark maroon bow, a small silver pendant necklace, black thigh-high stockings, and white platform heels. Both of her arms are fully mechanical—dark metallic gray prosthetic limbs with articulated joints and claw-like fingers. She is frozen in a powerful forward-leaning combat stance on cracked concrete amid rubble, one robotic hand thrust toward the camera, the other pulled back. The background is a ruined city at sunset: orange-red sky, silhouetted skyscrapers, one distant tower belching black smoke and fire, scattered debris and broken pipes on the ground. Dust motes float in the warm light. She holds the exact pose for a beat, then her eyes narrow with resolve as the young woman with a clear, determined mid-range voice (S1) says: <d>[English] You think you can just walk in and take what’s left?</d> Her mechanical fingers curl slightly. [Shot 2] At 00:04.200, the camera cuts to a medium shot tracking her as she pushes off her back foot and lunges forward through the rubble, robotic arms pumping. Debris scatters under her white heels. She continues without pause: <d>[English] I’ve already lost everything once.</d> Wind lifts strands of her blonde hair. [Shot 3] At 00:08.500, the camera cuts to a close-up of her face and extended right prosthetic hand as she plants her feet and swings the mechanical arm in a wide arc, palm open toward the unseen threat. Sparks flicker briefly at the wrist joints. Her expression hardens further while she finishes: <d>[English] Not again. Not while I’m still standing.</d> The camera slowly pushes in with small amplitude at slow speed on her determined eyes and the glowing sunset reflecting off the metal of her arm. [Shot 4] At 00:12.800, the shot holds on her full figure as she settles back into a ready stance similar to the opening, chest rising with controlled breaths, robotic hands half-open and ready, the ruined city and burning tower still framing her against the deepening orange sky. overall_soundscape: Distant low rumble of collapsing structures and a steady dry wind across the broken concrete. Metallic joint clicks and soft hydraulic hiss accompany every movement of her prosthetic arms. Her heels scrape and crunch over gravel and debris. Occasional sharp crackle of distant fire and falling masonry. non_diegetic_music: Driving mid-tempo electronic pulse layered with deep distorted synth bass and sparse metallic percussion hits that intensify during her lunge, then settle into a sustained low drone under the final stance.
Render by saehara151 on
Civitai
·
Prompt on BunchaPrompts
Flag
Settings
Size
672x896
Prompt
For the target video, at 0.00 seconds into the target video, <Picture 1> (from [Shot 1]) is fully referenced. integrated_multimodal_description: [Shot 1] Anime style. A woman with crimson hair in a loose ponytail, red eyes, white dress, black leather jacket, black belt, sheer black stockings, black boots. She stands on neon-lit steps under pink and blue light. Low-angle medium shot. Camera follows her as she steps forward. Her right hand brushes back a strand of hair. She holds eye contact with a smirk, lips parting slightly. Camera tilts up gently to follow her gaze. Shallow depth of field isolates her from the background. (S1) [Japanese] "Welcome to the show." overall_soundscape: Low hum of neon signs. Faint hiss of distant crowd. Crisp leather creak as she adjusts her jacket. Soft click of heels on polished steps. Gentle brush of fabric. Subtle pulse of stage lights. non_diegetic_music: Sultry neon-noir groove. Pulsing analog synth bass. Tight funk guitar stabs. Brushed drums with crisp hi-hats. Subtle electric piano chords. Smooth vocal hum. Medium intensity. Rhythmic.
Run Locally
polyhedron_minimax_h3_image_lora_v2__perfe8ct.safetensors
296 MB · SafeTensor · bf16
Downloads are served by Civitai. Some creators ask for a Civitai login first.