Model
.ai
tracking open ai models

Natural Face & Speech Minimax H3 v2.0 MH3

8520°
Natural Face & Speech Minimax H3 v2.0 MH3
Created by
AdaptiveVision
on Civitai
by AdaptiveVision - Natural Face & Speech Minimax H3 V2
Creator notes
Face and speech dynamics for Minimax H3 Full face muscle dynamics with clear English speech Trained on T2V but supports also I2V 720x1280 resolution with 15-30 steps Weight 0.6-0.8 is recommended Pro Tip: to prevent voice drift, use fixed seed and start your text with the same word (hey guys,) Prompt: she is speaking into a mic, she maintains eye contact with the viewer saying: 'text here'
2,110 downloads · 95 likes on Civitai
SHA256
40B1AF6134DDF8597E6B9BFED1B97FC782BAC6917E498F954EFFD34681AB881B
Ranking badge
Updated Oct 22, 2026
Natural Face & Speech Minimax H3 v2.0 MH3 on ModelTrends.ai: #169 Hottest AI model
Made this model? Show this badge on your page.
Paste it into your Hugging Face card or Civitai description.
Also ranked: #10 Hottest MiniMax H3 model · #10 Hottest MiniMax H3 LoRA
Not this creator?
Natural Face & Speech Minimax H3 v2.0 MH3 Popularity Trend
Heat score since 19 Sep 2026
Heat score
8,520°
This week
▲
9.1%
8,502° then, 8,520° now
This month
n/a
Wasn't released/tracked
Past year
n/a
Wasn't released/tracked
Natural Face & Speech Minimax H3 v2.0 MH3: Heat score since 19 Sep 20267,000°7,426°7,851°8,277°8,702°19 Sep26 Sep2 Oct9 Oct16 Oct22 Oct8,520°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.
+1,017°
Civitai downloads
2,110 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.
+297°
Likes
95 likes and thumbs up across the sources. A like weighs half a download, on the same log scale.
+106°
Adoption rate
About 25 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 8,520°, 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.
    2,110 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.
    +1,017°
    8,017°
    running total
  3. 3
    Add the likes.
    95 likes and thumbs up, at half the weight of a download.
    +297°
    8,314°
    running total
  4. 4
    Add the adoption rate.
    About 25 downloads a day right now, measured between two readings, at a quarter of the weight.
    +106°
    8,420°
    running total
  5. 5
    Family boost.
    An open, bounded nudge for LoRAs of a family this site is pushing forward right now.
    +100°
    8,520°
    running total
  6. =
    The heat score.
    Where this model stands on every list of this site. Nothing else adjusts it.
    8,520°
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.
Example render 1
Example render 2
Example render 3
Example render 4
Example render 5
NSFW render, blurred. Turn NSFW on to see it.
Example render 6
NSFW render, blurred. Turn NSFW on to see it.
Example render 7
NSFW render, blurred. Turn NSFW on to see it.
Example render 8
NSFW render, blurred. Turn NSFW on to see it.
Render by Gork71 on
Civitai
·
Prompt on BunchaPrompts
Flag
Settings
Size
1280x736
Prompt
A 2010s–2020s DC cinematic film, shot on digital IMAX with a 2.39:1 anamorphic widescreen aspect ratio. Justice League lounge room interior — brushed silver and gunmetal architecture, comfy sofas, lounge chairs, bookshelf, fireplace, vending machine. Cool white practicals in the ceiling grid, hard rim light, crushed blacks, faint anamorphic flare, fine digital grain. Formal, "giving a tour of the place" type atmosphere. integrated_multimodal_description: [Shot 1]: Static wide. Establishing shot of Justice League lounge room interior. Batman is standing in the lounge room, Supergirl and Superman are standing behind him several feet back, Flash is off to the side leaning against the wall. Superman in the classic blue-and-red cinematic suit, Batman in the textured black tactical cowl and cape, The Flash in the red-and-gold Speed Force suit, Supergirl in her iconic blue suit and red cape, hands at her sides, waiting. [Shot 2]: At 00:02:00, medium close-up on Batman, looking down at his thin slate, Supergirl and Superman behind his shoulder. "Okay, here are the sleeping arrangements. Only 3 rooms. So, Supergirl," he looks up dramatically from the thin slate "you're bunking with me Tuesdays and Thursdays” [Shot 3]: At 00:05:00, medium close-up on Supergirl. Her brow knits. (confused): “Wait, what?” [Shot 4]: At 00:06:00, medium shot on Batman as he spins around dramatically to face Supergirl, his cape sweeping behind him. BATMAN: “With Flash Mondays and Wednesdays." [Shot 5]: At 00:08:00, medium shot on Flash as he silently gives himself a small fist pump in celebration. [SHOT 6]: At 00:09:00, medium close-up on Supergirl. “Now just hold on a...” [SHOT 7]: At 00:10:00, medium close-up on Batman as he interrupts her. “And with Superman Fridays through to Sundays. [SHOT 8]: At 00:11:00, medium shot on Supergirl and Superman who both look disgusted. SUPERGIRL: “Okay, Gross! We're cousins." she says, throwing up her hands and glancing at Superman before turning back to focus on Batman. Meanwhile Superman maintains his look of disgust until he takes a step backwards and once out of Supergirl's eye line he drops the disgusted act and starts grinning, giving Batman a double thumbs up and a big wink. overall_soundscape: Quiet space-station hush — low air-circulation hum, crackling of fireplace, hum of vending machine, the sound of leather stretching with Flash's fist pump, the dramatic SWOOSH of Batman's cape, the subtle grunt of effort from Superman's air thrusts. Audio quality: Dry, close, enclosed-station recording. Voices are clean and unprocessed. non_diegetic_music: Light heroic orchestral underscore in early-2000s CW television style featuring strings and soft brass, present at low volume during opening shot, fading slightly when dialogue begins at 00:01:50. Cello plucks during comedic moments.
Render by AdaptiveVision on
Civitai
·
Prompt on BunchaPrompts
Flag
Settings
Steps
30
Size
544x960
Prompt
[Shot 1 | 00:00 - 00:04] Photorealistic 4K broadcast video, professional modern news studio with sleek backlit display panels and soft studio lighting. A confident 35-year-old male news anchor in a tailored dark navy suit and tie sits at a high-end anchor desk. The camera smoothly zooms in from a medium shot to a tight chest-up composition. As the broadcast begins, he maintains steady direct eye contact with the camera, offering a subtle professional smile, and articulates with precise, natural mouth movements: "Good evening. Welcome to tonight's live news broadcast. We start with breaking developments unfolding this hour." [Shot 2 | 00:04 - 00:08] Hard cut to a tight close-up shot focusing on his face, jawline, and natural expressions. He nods slightly for emphasis as he continues speaking, his mouth, teeth, and cheek muscles moving in tight, fluent sync with his vocal cadence: "Our top story tonight covers key updates from across the nation as events continue to progress."
Render by AdaptiveVision on
Civitai
·
Prompt on BunchaPrompts
Flag
Settings
Steps
30
Size
544x960
Prompt
[Shot 1 | 00:00 - 00:04] Ultra-realistic 4K Instagram reel aesthetic, soft flattering vanity lighting. A stylish 24-year-old woman with sleek dark hair tucked behind her ears sits in a minimalist beauty studio with ring light reflection in her eyes. Holding a sleek lip gloss wand near her face, she looks directly into the vertical camera lens with an inviting smile and speaks with clear articulation: "Hey guys! Today I'm sharing my absolute favorite technique for creating the ultimate glossy pout!" [Shot 2 | 00:04 - 00:08] Hard cut to an extreme close-up of her lips and jawline. She smoothly applies a dab of gloss precisely onto her cupid's bow, softly presses her lips together, and looks up into the camera while articulating naturally and in sync: "Just press a tiny touch right on the cupid's bow, and look at that incredible shine!"
Render by AdaptiveVision on
Civitai
·
Prompt on BunchaPrompts
Flag
Settings
Steps
30
Size
544x960
Prompt
[Shot 1 | 00:00 - 00:04] Photorealistic 4K cinematic video, natural lighting. A cheerful 25-year-old woman with curly brown hair, wearing a casual denim jacket, stands in a brightly lit, modern artisanal pizzeria. She holds a steaming, cheese-pulling slice of pepperoni pizza on a wooden board up to the camera. The camera maintains a stable medium shot with smooth handheld stabilization, tracking forward slightly. Looking directly into the camera lens with a wide, enthusiastic smile, she says: "Oh my goodness, look at this slice! You won't believe how incredible this smells!" [Shot 2 | 00:04 - 00:08] Hard cut to a close-up angle centered on her face and the slice. She takes a generous bite, steam rising from the crust as cheese pulls away cleanly. Her eyes widen in genuine delight, chewing with satisfaction and giving an approving nod toward the lens. Her mouth movements are natural and in precise sync with her spoken dialogue: "Mmm! That crispy crust and spicy pepperoni combo is absolute perfection!"
Settings
Sampler
LCM
Steps
8
Size
1728x1056
Prompt
minimax
Settings
Size
1280x720
Prompt
subject_definitions: <Subject 1> is the voluptuous Japanese woman from <Picture 1>, featuring long dark hair tied in a ponytail, large breasts, and wearing only a white fabric apron over her bare skin. <Subject 2> is the domestic kitchen environment from <Picture 1>, characterized by cluttered countertops, tiled walls, wooden cabinets, and warm incandescent lighting. summary: [reference generation] The target video features <Subject 1> cooking in <Subject 2>. The scene is styled as a low-resolution VHS screencap with grainy textures and nostalgic color bleeding. A narrator's voiceover provides an internal monologue for the woman. retention_analysis: <Subject 1> (appears in [Shot 1]): fully_preserved - the woman's appearance, hair style, large breasts, and white apron are retained from <Picture 1>. <Subject 2> (appears in [Shot 1]): fully_preserved - the kitchen layout, cluttered counter objects, tiling, and warm lighting are retained from <Picture 1>. detailed_description: The target video is presented in a grainy, low-resolution VHS screencap style typical of Japanese adult videos, featuring heavy color bleeding, slight motion blur, and a desaturated nostalgic aesthetic. [Shot 1] A medium shot captures <Subject 1>, the voluptuous Japanese woman with long hair in a ponytail and large breasts, as she stands in the middle of <Subject 2>, the cluttered domestic kitchen. She is dressed only in a white fabric apron over her bare skin. The warm incandescent light from above casts soft highlights on her shoulders and the surrounding tiled walls. <Subject 1> holds a wooden spatula in her right hand, gently stirring food inside a black frying pan placed on a lit gas stove burner. Her head is slightly tilted down as she focuses on the cooking process. A narrator's voice (S1), representing the woman's internal thoughts, provides a somber narration over the scene. The narrator says, <d>[Japanese] まで数日、私の人生は退屈でした。</d> The camera maintains a steady position with very slight handheld jitter to simulate an old home video recording. As she stirs, the steam from the pan rises faintly into the warm air, and her movements are slow and deliberate, emphasizing the domestic atmosphere of the scene. overall_soundscape: The low hum of a gas stove burner and the faint sizzling sound of food cooking in the pan provide the primary diegetic audio throughout the shot. non_diegetic_music: N/A
Settings
Size
768x1344
Prompt
subject_definitions: <Subject 1> is the elf named Diana, whose appearance is defined by <Picture 1>, <Picture 2>, <Picture 3>, and <Picture 4>. She has long, wavy blonde hair with orange highlights, pointed ears, blue eyes, and wears a white top, dark utility belt, and a dark blue cape with red lining. Her expression is consistently cheerful and expectant, and she is shown in full-body, medium, and close-up shots. summary: [reference generation] The target video features <Subject 1> (Diana) waiting expectantly for a response, then walking side by side with a Hero through a busy guild hall toward a bar. The scene uses <Picture 1> as the full-body reference for Diana’s pose and outfit, <Picture 2> as the medium shot reference for her facial expression and upper-body framing, and <Picture 3> and <Picture 4> as close-up references for her eye expression and hair detail. The camera follows Diana at eye level as she walks, and subtitles appear as the Hero speaks. retention_analysis: <Subject 1> (appears in [Shot 1], [Shot 2]): fully_preserved - Diana’s long blonde hair with orange highlights, pointed ears, blue eyes, white top, dark utility belt, and dark blue cape with red lining are retained. Her cheerful and expectant expression is preserved across all shots. The full-body pose from <Picture 1> is used in [Shot 1], the medium shot framing and facial expression from <Picture 2> are used in [Shot 2], and the close-up eye and hair detail from <Picture 3> and <Picture 4> are referenced for emotional continuity. detailed_description: The target video is in a cinematic, fantasy-adventure style with warm, ambient lighting and a slightly desaturated color palette to emphasize the mood of anticipation and camaraderie. [Shot 1] The scene opens with a full-body shot of <Subject 1> (Diana), standing in a neutral pose with a bright, expectant smile. Her long, wavy blonde hair with orange highlights flows naturally, her pointed ears are visible, and she wears her white top, dark utility belt, and dark blue cape with red lining. The background is softly blurred to focus on her expression. Subtitles appear as the Hero speaks: <d>[English] Diana, is it? You’re wasted standing here. Come. Let’s talk over a drink.</d> The camera remains static, capturing her posture and expression as she waits for a response. [Shot 2] At 00:15.000, the shot cuts to a medium tracking shot following Diana and the Hero as they walk side by side through a busy guild hall toward an attached bar. The camera stays at eye level, smoothly tracking their movement. Diana is animated, chatting excitedly with the Hero. Her expression remains cheerful and engaged, and her hair and clothing are consistent with the reference images. The background shows bustling activity with other patrons and guild members, but the focus remains on Diana’s interaction. The camera movement is steady and natural, following their pace without abrupt cuts. The scene ends as they approach the bar. overall_soundscape: Ambient background noise of a busy guild hall continues throughout the scene, including distant chatter, footsteps, and clinking of mugs. non_diegetic_music: N/A
Settings
Size
768x1376
Prompt
integrated_multimodal_description: [Shot 1] Live-action, cinematic, a vertical 9:16 medium shot of a beautiful woman standing in an intimate, softly lit bedroom. She is completely nude, facing the camera and displaying her breasts prominently as she maintains direct eye contact with a seductive expression. The camera holds a static shot to maintain focus on her upper body and facial expressions. With a sultry, beckoning tone (S1), she leans slightly forward and says: <d>[English] My darling, you are forbidden to cum until I tell you. Take your cock with one hand and with the other one rub your balls.</d> She gestures with her hands in a slow, rhythmic motion as if guiding the viewer's actions while maintaining a consistent gaze throughout the 10-second duration. overall_soundscape: Soft breathing from the woman and the subtle rustle of fabric from nearby bedding. The room has a quiet, intimate ambience. non_diegetic_music: N/A
Run Locally
nafasp_h3_lora_v2_500.safetensors
296 MB · SafeTensor · bf16
Downloads are served by Civitai. Some creators ask for a Civitai login first.