Hunyuan Image 3.0 is Tencent's open-source image model — a native multimodal, unified autoregressive model with 80 billion parameters (about 13B activated) across a 64-expert mixture, and the largest open-source image model at launch. It rewards two habits: writing prompts as ordered, natural-language descriptions, and putting any in-image words in double quotes. These 18+ templates bake both habits in. Each one is a skeleton with [bracketed placeholders] — swap the brackets for your own specifics and generate.

How to use these templates: copy a template, replace every [bracket] with concrete words, keep any "[QUOTED TEXT]" as your exact in-image copy, then pick the aspect ratio from the Settings line and generate. Do not rewrite the sentence order — that ordering is what Hunyuan reads best. The master formula behind the photo templates:

[subject + appearance] + [action/pose] + [setting] + [lighting] + [camera & lens] + [style], with any in-image words as "[EXACT WORDS]"

Keep the Hunyuan Image cheat sheet open for the modifier tables, browse the best Hunyuan Image prompts roundup for finished examples, and read the photorealism guide for the lighting and lens language behind the photo templates.

Advertisement

Photoreal scenes & portraits

These follow the master formula. Set the light and time of day, then the lens. Hunyuan's world knowledge fills real-world detail, so name a place or material and swap the rest.

1. Photoreal scene / landscape

A [subject / place] at [time of day], [key detail or atmospheric element such as mist, reflections, or spray], [secondary element for scale], [color mood], photorealistic [landscape / architectural] photography with [depth of field], [lighting direction and quality], shot on a [focal length] lens.

Settings: Hunyuan Image 3.0 · [3:2 or 16:9] · 2K

Fill in: subject/place, time of day, atmospheric detail, scale element, color mood, depth of field, lighting, focal length.

2. Environmental portrait

A [person: age, features, wardrobe] with [distinguishing detail], medium environmental portrait as they [action] in [their setting with background props], lit by [light source and quality with fill], natural documentary photography with honest color and fine skin texture, [depth of field], shot on a [35mm or 50mm] lens.

Settings: Hunyuan Image 3.0 · 4:5 · 2K

Fill in: person and wardrobe, distinguishing detail, action, setting and props, light source, depth of field, focal length.

3. Studio headshot / beauty portrait

A [person: age, features, hair] with [skin/expression detail], [shot type: straight-on beauty headshot / three-quarter portrait] [expression], against a seamless [backdrop color] backdrop, lit with [lighting setup, e.g. high-key beauty dish or low-key chiaroscuro] for [effect], [commercial / cinematic] retouched look with true-to-life color, shot on a [85mm or 100mm] lens.

Settings: Hunyuan Image 3.0 · [1:1 or 4:5] · 2K

Fill in: person and hair, expression, shot type, backdrop color, lighting setup and effect, look, focal length.

Product & e-commerce

Clean studio light, a defined surface, and a macro lens make materials read premium. Center the product, keep the background simple, and swap the object and palette.

4. Product hero shot

A [product with material and color] centered [three-quarter / front] view on a [surface] in a minimalist studio, [one supporting prop or effect], lit with a soft [three-point / single softbox] setup and a subtle rim light to define the edges, photorealistic product-studio style with crisp reflections and a seamless [backdrop color] backdrop, shot on a 100mm macro lens with shallow depth of field.

Settings: Hunyuan Image 3.0 · [4:5 or 1:1] · 2K

Fill in: product and material, view, surface, supporting prop, lighting setup, backdrop color.

5. Dynamic / levitation product shot

A single [product] floating mid-air at a dynamic angle with [motion detail, e.g. trailing laces or splash], centered composition against a bold gradient studio backdrop of [color one] to [color two], lit with punchy directional light and a bright rim to separate it from the background, high-contrast commercial product style with sharp texture, shot on a [35mm or 50mm] lens.

Settings: Hunyuan Image 3.0 · 1:1 · 2K

Fill in: product, motion detail, two gradient colors, focal length.

6. Flat-lay / lifestyle product scene

A [product] [lying flat / placed] on a [surface], [top-down flat-lay / three-quarter] view surrounded by [two or three complementary props], lit with soft [overcast diffusion / warm low light] for even natural light and gentle shadows, [warm editorial / cozy lifestyle] brand style with rich texture and honest color, shot on a [35mm or 50mm] lens.

Settings: Hunyuan Image 3.0 · 4:5 · 2K

Fill in: product, surface, view, props, lighting, brand style, focal length.

Advertisement

Text, posters & packaging

This is a signature strength. Put the exact copy in double quotes, name the font, describe placement, and keep each string short. Hunyuan renders legible English and Chinese, so bilingual layouts are a real edge.

7. Poster with headline

A [poster type] with the headline "[HEADLINE]" in a [font style, e.g. heavy condensed sans-serif] across the [position] and the subtitle "[SUBTITLE]" in a [smaller font] beneath it. Below the type, [central illustration or photo scene], [art style and palette]. Headline [placement], everything sharp and correctly spelled.

Settings: Hunyuan Image 3.0 · 2:3 · 2K

Fill in: poster type, "[HEADLINE]", font, position, "[SUBTITLE]", scene, art style, placement.

8. Packaging label

A front-facing product shot of a [material and color] [package] with a printed label reading "[BRAND NAME]" in a [font], the line "[SUBLINE / VARIANT]" in small caps beneath it, and a simple [line-art icon] above the name. [Label palette] on a [package color], soft studio light, photorealistic packaging mockup with legible, correctly spelled type.

Settings: Hunyuan Image 3.0 · 4:5 · 2K

Fill in: package material, "[BRAND NAME]", font, "[SUBLINE]", icon, label palette, package color.

9. Bilingual flyer (English + Chinese)

A clean [event / promo] flyer with the English headline "[ENGLISH HEADLINE]" in a bold [font] and the Chinese subtitle "[中文标题]" directly below in a matching weight, plus a footer line "[DATE - TIME]". [Festive flat illustration or theme] on a [background color], generous spacing, all text sharp and correctly rendered in both languages.

Settings: Hunyuan Image 3.0 · 4:5 · 2K

Fill in: flyer type, "[ENGLISH HEADLINE]", font, "[中文标题]", "[DATE - TIME]", illustration theme, background color.

Logos & brand marks

Ask for flat, simple marks on a solid background, keep the palette to one or two colors, and quote the brand name when you want lettering. Swap the symbol, name, and style.

10. Wordmark logo

A flat vector logo for a [business type], a simplified [symbol] above the wordmark "[BRAND NAME]" in a [font style], one [color] on a solid [background color] background, centered and symmetrical, minimal graphic style with no gradients, shadows, or 3D. Keep the lettering crisp and evenly spaced.

Settings: Hunyuan Image 3.0 · 1:1 · 2K

Fill in: business type, symbol, "[BRAND NAME]", font, color, background color.

11. Abstract symbol / app mark

A flat vector logo for a [business type], an abstract geometric mark of [shape concept, e.g. two overlapping triangles suggesting upward motion], single [color] on a solid [background color] background, centered, bold and minimal with clean edges and no text, no gradient or shadow, modern app-icon feel.

Settings: Hunyuan Image 3.0 · 1:1 · 2K

Fill in: business type, shape concept, color, background color.

12. Vintage badge logo

A vintage circular badge logo for a [business type] with the curved text "[TOP TEXT]" around the top edge and "[BOTTOM TEXT]" along the bottom, a simplified [central emblem] in the center, flat two-tone design in [color one] and [color two], retro monoline style on a solid background, all lettering sharp and legible.

Settings: Hunyuan Image 3.0 · 1:1 · 2K

Fill in: business type, "[TOP TEXT]", "[BOTTOM TEXT]", emblem, two colors.

Anime, ink-wash & illustration

Hunyuan handles anime, ink-wash, comic, and flat-vector styles well. Name the style, keep the character description tight, and use a plain background for reusable sheets.

13. Anime key art

Dramatic anime key art of a [character: age, hair, wardrobe] [action/pose] in [setting] at [time of day], [dynamic element, e.g. wind, glowing weapon, lens flare], [camera angle] composition, vibrant cel-shaded anime style with clean linework and rich color, cinematic lighting.

Settings: Hunyuan Image 3.0 · 16:9 · 2K

Fill in: character and wardrobe, action, setting, time of day, dynamic element, camera angle.

14. Character sheet / mascot

A [style, e.g. flat-vector or watercolor-and-ink] character sheet for a [character] named [name], [front, side, and three-quarter views / single front-facing pose] on one sheet, wearing [distinguishing outfit or accessory], plain [background color] background, keep the proportions, [accessory], and color palette identical across all poses.

Settings: Hunyuan Image 3.0 · [16:9 or 1:1] · 2K

Fill in: style, character, name, views, outfit/accessory, background color.

15. Traditional ink-wash scene

A traditional Chinese ink-wash painting of a [subject] [action] in [setting], distant [background element] fading into pale washes of grey and black, a single splash of [accent color] on [detail], expressive loose brushwork with visible rice-paper texture and generous negative space, [mood], minimal and poetic.

Settings: Hunyuan Image 3.0 · 3:4 · 2K

Fill in: subject, action, setting, background element, accent color, detail, mood.

World scenes & Instruct edits

Hunyuan's native reasoning shines when a prompt leans on real facts — history, landmarks, science. The Instruct variant edits existing images and fuses references; attach the image(s) and say what to keep identical.

16. World-knowledge / historical scene

A richly detailed scene of [named historical period, landmark, or cultural event], [key figures or subjects with authentic detail], [environment and architecture], [lighting and time of day], historically informed [costume / setting] detail, [painterly realism / photorealistic] style with fine detail and rich color.

Settings: Hunyuan Image 3.0 · 16:9 · 2K

Fill in: named period or landmark, figures, environment, lighting, detail focus, style.

17. Labeled infographic / diagram

A clean educational infographic titled "[TITLE]" at the top, showing labeled stages — "[LABEL 1]", "[LABEL 2]", "[LABEL 3]", and "[LABEL 4]" — around a stylized [subject illustration], with thin arrows connecting each stage. Flat modern vector style, [color palette], all labels sharp and correctly spelled.

Settings: Hunyuan Image 3.0 · 4:3 · 2K

Fill in: "[TITLE]", the four "[LABELS]", subject illustration, color palette.

18. Background swap edit (Instruct)

Replace the [current background] behind the [subject] with [new background at time of day], keep the subject, their pose, clothing, hair, and lighting exactly the same, match the new background's [color temperature] with a subtle rim so they sit naturally in the scene, photorealistic result with clean edges around the hair.

Settings: Hunyuan Image 3.0-Instruct · edit · 1 reference image · keep subject identical

Fill in: current background, subject, new background, color temperature.

19. Two-image fusion (Instruct)

Combine the two references into one photorealistic scene: place the [element from image 1] on the [element from image 2], matching the perspective and the [lighting] of image 2, add a natural contact shadow, keep the [image-1 element]'s shape, [label / detail], and color exactly as in image 1, [premium product-photography / natural] look.

Settings: Hunyuan Image 3.0-Instruct · fusion · 2 reference images · keep image-1 element identical

Fill in: element from image 1, element from image 2, lighting, detail to preserve, look.

20. Style transfer edit (Instruct)

Repaint this photograph as a [target medium, e.g. traditional ink-wash or oil painting], keep the overall composition, [key structural elements], and the position of [main subject], translate the tones into [palette and brushwork], add [texture or negative-space note], keep it recognizable as the same scene.

Settings: Hunyuan Image 3.0-Instruct · edit · 1 reference image · preserve composition

Fill in: target medium, structural elements, main subject, palette and brushwork, texture note.

Frequently Asked Questions

How do I use a fill-in-the-blank Hunyuan prompt template?

Copy a template, then replace every bracketed placeholder with your own specifics before you generate. A template like [subject + appearance], [action] in [setting], [lighting], [lens], [style] becomes a real prompt once you swap each bracket for concrete words, for example a chef plating a dish in a warm kitchen, soft window light, 50mm lens, photorealistic. Leave the sentence structure and the descriptive order intact — that ordering is what makes Hunyuan Image 3.0 read the prompt cleanly — and only change the words inside the brackets. Then pick the aspect ratio and resolution from the Settings line and generate.

What do the [brackets] in these templates mean?

The brackets are blanks you fill in. Each one names the kind of detail that belongs there: [subject + appearance] is who or what plus how it looks, [action/pose] is what they are doing, [setting] is the place, [lighting] is the light quality and direction, [camera & lens] is the shot type and focal length, and [style] is the medium or look. Words in double quotes like "[HEADLINE]" are the exact letters you want printed inside the image, so replace them with your real copy and keep the quotes.

What aspect ratios work best for these templates?

Match the ratio to the use. Use 1:1 for logos, app icons, and social squares; 4:5 or 2:3 for portraits, posters, and packaging; 3:2 or 16:9 for landscapes, key art, and banners; and 3:4 or 9:16 for tall mobile scenes. Aspect ratio and resolution are interface or API settings in Hunyuan Image 3.0, not inline flags — there is no --ar parameter. Each template's Settings line suggests a ratio; select it and a size up to 2048x2048 (2K) in your tool.

How do I get the text right in the poster and logo templates?

Keep the exact words inside double quotes, keep each string short (roughly two to ten words), name the font style, and say where the text sits. Hunyuan Image 3.0 renders both English and Chinese characters accurately, so bilingual posters and packaging are a genuine strength. If a word comes out wrong, shorten it, isolate it as its own quoted element, and regenerate rather than adding more copy for the model to juggle.

How do the Instruct edit templates work?

The edit and fusion templates are for HunyuanImage 3.0-Instruct, the instruction-tuned variant that edits existing images and fuses references. Attach the reference image or images, fill in the bracketed instruction, and name what to keep identical. For a background swap you keep the subject; for a two-image fusion you say which element comes from which reference. The Settings line under each edit template notes how many reference images to attach and what to preserve.

Can I make the templates shorter or longer?

Yes. Hunyuan Image 3.0 handles both. Thanks to its world-knowledge reasoning, a short filled-in template still resolves into a detailed, plausible scene, so you can trim brackets you do not care about. It also understands very long prompts of well over a thousand characters, so when you need tight control you can expand a template with extra detail on wardrobe, props, color, and mood, and it will honor the specifics instead of dropping them.

What is Hunyuan Image 3.0 and who makes it?

Hunyuan Image 3.0 is Tencent's open-source text-to-image model, released under an Apache 2.0 license for free commercial use. It is a native multimodal, unified autoregressive model with 80 billion total parameters (about 13 billion activated per token) across a 64-expert mixture-of-experts, which makes it the largest open-source image model at launch. Its strengths are world-knowledge reasoning, best-in-class bilingual English and Chinese in-image text, a wide style range, and strong Asian cultural authenticity. You can run it from the weights on Hugging Face and GitHub, use hunyuan.tencent.com, or reach it through APIs like Tencent Cloud and WaveSpeed.

Advertisement