<h2 id="simple-minimax-h3-image-workflow-(text-to-image)">Simple MiniMax H3-Image workflow (text to image)</h2><p>A few people asked about blurry or weird eyes with the official ComfyUI template. The template has two problems: it renders at only 4 MP, and it uses euler. H3-Image works with big 32×32 px patches, so at 4 MP the eyes in a full-body shot are only a few patches wide and turn to mush. Euler also gives noisier skin and makes faces look older.</p><p>This is a stripped-down workflow that fixes both:</p><p>16 MP output (a half-size copy is saved for viewing)</p><p>SamplerLCM (s_noise 1.10 → 0.90), simple, 6 steps, cfg 1</p><p>decodes with the P-H3 VAE for a bit more detail</p><h3 id="only-core-nodes-comfyui-0.45-or-newer.">Only core nodes, ComfyUI 0.45 or newer.</h3><p>The download links are in the note inside the workflow. If you want to use a LoRA, there’s an optional slot: pick your LoRA, select the node and press Ctrl+B. The same goes for the full 16 MP save. You can also drag the example image into ComfyUI to get the workflow with its exact settings.</p><h3 id="reference-image-ready!">Reference Image Ready!</h3><p>Add a Load Image node and plug it into image_0 on the conditioning node. That’s it, the VAE is already connected. In your prompt the images are &lt;Picture 1&gt;, &lt;Picture 2&gt; and so on. Picture 1 is the one that gets edited, or you can just use it as a reference for a new scene. There’s a short note in the workflow with examples.</p><p></p><h3 id="made-for-my-blender-style-lora">Made for my Blender Style Lora</h3><p></p>

H3I Fixed! Extremely Simple 16MP workflow T2I (I2I Ready)

MiniMax_H3-Image_simple_T2I.json

T2I v1.0

<h1 id="3d-render-style-for-minimax-h3-image">3D Render Style for MiniMax H3-Image</h1><p>A style LoRA trained on a set of about 400 3D character renders. It gives H3-Image the look of polished 3D renders: CG skin and materials, render-style lighting and proportions.</p><h3 id="how-to-use">How to use</h3><p>- Start your prompt with A 3d render.</p><p>- Strength 1.0 is the default. You can go down to about 0.8 for a lighter effect, but that is less tested.</p><p>- Tested in ComfyUI with the distilled 6-step H3-Image at up to 16 MP.</p><h3 id="good-to-know">Good to know</h3><p>- The LoRA pushes a strong athletic, curvy body type (defined abs, large proportions), even when the prompt asks for a slimmer figure.</p><p>- Every preview image has its full prompt and workflow in the metadata.</p><h3 id="example-prompt">Example prompt</h3><p>&gt; A 3d render. Close selfie shot of an adult woman taken from her own outstretched right arm, which reaches toward the camera at the right edge of the frame and is strongly foreshortened, framed from the top of her head to her hips, her body turned slightly to the left. She has fair skin with freckles across her nose and shoulders, flushed cheeks and a light sheen of sweat, bright green eyes with smoky eyeshadow, full pouting lips and long copper-red hair pulled back into a messy high ponytail with loose strands stuck to her damp face. She has an athletic figure with large round breasts, toned arms, defined abs and a narrow waist. She is fully nude; her bare breasts with pink nipples are fully visible, and a few glossy streaks of white semen run across her collarbone and upper chest. Her left hand is raised beside her face making a peace sign with two fingers, her nails painted dark red, and she gives the camera a smug, satisfied look with one eyebrow slightly raised. A semi-transparent dark grey caption bar runs straight across the lower third of the image with the short white text \"Mission Accomplished!\" centered in it in a simple sans-serif font. The background is a dark, softly lit grey room that is completely out of focus. The air is filled with a thin haze that softens the light, giving the image a dreamy, misty glow, lifted milky shadows and a gentle bloom on her shiny skin; soft frontal light from the phone side lights her face and chest, and the corners of the frame fall off into soft darkness.</p><h3></h3><p></p>

Files with SHA256: 8ff77b39e2191d4510bbdbab7f4dfc246c5e39b11eda9d85b48037b8e975dc77