Tencent's image model on Pixel Dojo
Hunyuan Image 3 AI Image Generator
Hunyuan Image 3 turns long, detailed prompts into rich photoreal scenes, placing every person, prop and light you describe in one frame.
Hunyuan Image 3, also known as HunyuanImage 3.0, is Tencent's 80 billion parameter image model. Make street scenes, portraits, food spreads, architecture, concept art and lettered posters from prompts as long as a full paragraph, in seven aspect ratios, for 2 credits an image. We made every example on this page on Pixel Dojo, and each one comes with the exact prompt behind it.
2 credits ($0.13) per image. Included in every plan, from $10/month.
Already a member? Sign in
By Brian Lovett · Updated
Hunyuan Image 3 examples and prompts
Six images we made with Hunyuan Image 3 at 2 credits each, five in Quality mode and the cliff library in Fast mode. Each prompt reads like a shot list, so reuse it as is or swap in your own subject. Tap Use this prompt to start from any of them.
Architecture at 16:9. Concrete terraces, a curving stair and one red coat for scale, lit by shafts of golden light. Swap in your own building or coastline.
Wide cinematic photograph of a vast brutalist library carved into a sea cliff at golden hour: tiers of concrete reading terraces stepping down toward the ocean, a spiral staircase of pale stone, long shafts of warm light cutting through dust, a single figure in a red coat crossing a narrow bridge for scale, waves breaking far below, deep shadows defining the angular geometry, atmospheric haze, painterly cinematic photography.
Concept art at 16:9. Houses, gardens, a lighthouse, kites, boats and birds from one prompt, all in scale. Start here for game and film worlds.
Epic concept art of a floating market city built on the shells of three colossal sea turtles swimming through a misty turquoise ocean at dawn: stilted wooden houses, terraced gardens, laundry lines and paper kites, a small lighthouse on the lead turtle, rowing boats ferrying crates between the shells, flocks of white birds, warm sunrise light glowing through the haze, hyper detailed painterly style with a sense of enormous scale.
Food photography at 4:3. Seven dishes, mint tea and two hands tearing flatbread, every item distinct under hard afternoon sun.
Overhead photograph of a generous mezze table on cobalt blue tiles: a swirl of hummus with paprika and a pool of olive oil, bright green tabbouleh, charred eggplant scattered with pomegranate seeds, a bowl of glossy olives, golden falafel, pickled pink turnips, warm puffed flatbread being torn by two hands, small glasses of mint tea, scattered parsley, hard late afternoon sun with crisp shadows, saturated appetizing color, every dish distinct.
Editorial portrait at 3:4. Window light rakes across amber varnish, wood grain, a white beard and working hands. Start from this prompt for any craftsperson portrait.
Photoreal editorial portrait of a luthier in his seventies with a trimmed white beard and wire rimmed glasses, holding a newly finished violin up to a tall workshop window to inspect the amber varnish, his workshop crowded with hanging cellos, clamps, curled wood shavings and brass tools, soft directional window light raking across spruce, maple and skin, calm focused expression, warm muted palette, hyper detailed textures.
Street photography at 3:2. Fish scales, crushed ice, wet cobblestones and lantern glow in one candid frame with two people and a market full of props.
Photoreal street photograph at blue hour in a narrow old town alley market just after rain: a fishmonger in a yellow rubber apron lifts a silver sea bream from a bed of crushed ice, a cyclist with a wicker basket of lemons squeezes past, paper lanterns and a red neon sign glow overhead, steam rises from a dumpling stall, wet cobblestones mirror every light, candid documentary feel, rich natural color, fine detail on scales, ice and cloth.
Travel poster at 2:3. The headline and the bottom line print exactly in flat lithograph colors. Put your own words in quotes and add "the text must be exact".
Vintage travel poster for a fictional mountain railway: the bold headline "ALPINE EXPRESS" in large cream capitals across the top, a red electric train crossing a curved stone viaduct over a deep green valley, snowy peaks glowing pink at sunrise behind it, flat lithograph colors with subtle paper texture, the small line "DAILY FROM LUCERNE" at the bottom, the text must be exact.
Hunyuan Image 3 Fast mode and Quality mode, same 2 credits
Fast mode is the default, and Quality mode spends extra steps on edges and fine texture. Both cost 2 credits. We also ran three of the showcase prompts in Fast mode, so you can see how close the two modes come.
Fast mode at 4:3. The same mezze prompt as the Quality tile above, with the flatbread torn on the right.
Overhead photograph of a generous mezze table on cobalt blue tiles: a swirl of hummus with paprika and a pool of olive oil, bright green tabbouleh, charred eggplant scattered with pomegranate seeds, a bowl of glossy olives, golden falafel, pickled pink turnips, warm puffed flatbread being torn by two hands, small glasses of mint tea, scattered parsley, hard late afternoon sun with crisp shadows, saturated appetizing color, every dish distinct.
Fast mode at 16:9. Two turtle islands, kites, boats and a lighthouse from the same concept art prompt.
Epic concept art of a floating market city built on the shells of three colossal sea turtles swimming through a misty turquoise ocean at dawn: stilted wooden houses, terraced gardens, laundry lines and paper kites, a small lighthouse on the lead turtle, rowing boats ferrying crates between the shells, flocks of white birds, warm sunrise light glowing through the haze, hyper detailed painterly style with a sense of enormous scale.
Fast mode at 3:2. Lanterns, neon, steam and the lemon basket from the alley market prompt.
Photoreal street photograph at blue hour in a narrow old town alley market just after rain: a fishmonger in a yellow rubber apron lifts a silver sea bream from a bed of crushed ice, a cyclist with a wicker basket of lemons squeezes past, paper lanterns and a red neon sign glow overhead, steam rises from a dumpling stall, wet cobblestones mirror every light, candid documentary feel, rich natural color, fine detail on scales, ice and cloth.
- Write the whole scene. Hunyuan Image 3 reads long prompts: our bookshop cafe prompt below lists a dozen details, and every one is in the picture.
- Pair a structure with a light. "A brutalist library carved into a sea cliff at golden hour" and "a narrow old town alley market at blue hour" each gave Hunyuan a clear target.
- Add one figure for scale. A single person in a red coat on a bridge sells the size of the cliff library.
- Name materials and textures: spruce, maple and amber varnish, crushed ice and fish scales, cobalt tiles. Hunyuan renders surface detail you can almost touch.
- Quote short lines of text and add "the text must be exact". Our ALPINE EXPRESS poster printed its headline and bottom line as written.
- Fast and Quality cost the same 2 credits. Explore in Fast, then switch Mode to Quality for the image you keep.
- Pick the frame first: 16:9 for landscapes and architecture, 3:4 for portraits, 4:3 for food, 2:3 for posters.
Hunyuan Image 3 vs Seedream 5 on a 926 character prompt
We gave both models the same long bookshop cafe prompt at 3:2, with a dozen details to place: a ginger cat asleep on atlases, a reader in a moss green cardigan, a barista pouring latte art, a chalkboard reading SOUP OF THE DAY: PUMPKIN, a spiral staircase, fairy lights, a globe, a fern and a red bicycle in the rain. Both placed every one and spelled the chalkboard exactly. Hunyuan Image 3 painted a warm, crowded storybook frame with every prop up front for 2 credits. Seedream 5 Pro made a quieter, more photographic take for 2.5 credits.
Same prompt: A cozy bookshop cafe on a rainy autumn evening, seen from just inside the front door at eye level. On the left, floor to ceiling walnut shelves with a rolling library ladder on a brass rail. A ginger cat sleeps curled on a stack of old atlases on a round table. An elderly man in a moss green cardigan reads a folded newspaper under a brass lamp with a green glass shade. Behind the counter, a barista with a shaved head and a denim apron pours a latte art leaf into a white cup. A chalkboard on the counter reads "SOUP OF THE DAY: PUMPKIN". A black iron spiral staircase climbs to a mezzanine strung with warm fairy lights. An antique globe sits on a side table next to a potted fern. Rain streaks the big front window, and outside a red bicycle leans against a lamp post under a blurred street light. Warm tungsten interior against cool blue dusk outside, photoreal, 35mm lens, soft film grain, every detail clearly visible.
Hunyuan Image 3Hunyuan Image 3, Fast mode at 3:2. Every prop sits up front in warm tungsten light, from the sleeping cat to the globe and the red bicycle.
A cozy bookshop cafe on a rainy autumn evening, seen from just inside the front door at eye level. On the left, floor to ceiling walnut shelves with a rolling library ladder on a brass rail. A ginger cat sleeps curled on a stack of old atlases on a round table. An elderly man in a moss green cardigan reads a folded newspaper under a brass lamp with a green glass shade. Behind the counter, a barista with a shaved head and a denim apron pours a latte art leaf into a white cup. A chalkboard on the counter reads "SOUP OF THE DAY: PUMPKIN". A black iron spiral staircase climbs to a mezzanine strung with warm fairy lights. An antique globe sits on a side table next to a potted fern. Rain streaks the big front window, and outside a red bicycle leans against a lamp post under a blurred street light. Warm tungsten interior against cool blue dusk outside, photoreal, 35mm lens, soft film grain, every detail clearly visible.
Seedream 5Seedream 5 Pro at 3:2. A calm, photographic room with the same details set deeper into the space.
A cozy bookshop cafe on a rainy autumn evening, seen from just inside the front door at eye level. On the left, floor to ceiling walnut shelves with a rolling library ladder on a brass rail. A ginger cat sleeps curled on a stack of old atlases on a round table. An elderly man in a moss green cardigan reads a folded newspaper under a brass lamp with a green glass shade. Behind the counter, a barista with a shaved head and a denim apron pours a latte art leaf into a white cup. A chalkboard on the counter reads "SOUP OF THE DAY: PUMPKIN". A black iron spiral staircase climbs to a mezzanine strung with warm fairy lights. An antique globe sits on a side table next to a potted fern. Rain streaks the big front window, and outside a red bicycle leans against a lamp post under a blurred street light. Warm tungsten interior against cool blue dusk outside, photoreal, 35mm lens, soft film grain, every detail clearly visible.
| Hunyuan Image 3 | Seedream 5 | |
|---|---|---|
| Credits for this image | 2 | 2.5 |
| Details from the prompt | All twelve in frame | All twelve in frame |
| Chalkboard text | Spelled exactly | Spelled exactly |
| Look | Warm, dense storybook frame, props up front | Quiet and photographic, a deeper room |
What Hunyuan Image 3 does best
Long, detailed prompts
Describe a whole scene in a paragraph. Hunyuan Image 3 keeps track of people, props, light and weather, as in our 926 character bookshop cafe.
Photoreal people and places
Natural faces and hands, wet cobblestones, varnished wood and drifting steam, from the luthier portrait to the alley market.
Food and materials
Distinct dishes, glossy olive oil, crushed ice, fish scales and cobalt tiles, each with its own texture.
Architecture and concept art
Huge structures at believable scale, with a small figure or boat to sell the size, in painterly or photographic styles.
Short lettering
Quoted headlines and signs, like ALPINE EXPRESS on our travel poster and SOUP OF THE DAY: PUMPKIN on the cafe chalkboard.
Seven aspect ratios
1:1, 16:9, 9:16, 4:3, 3:4, 3:2 and 2:3, all at 2 credits. Our files measured 1280 x 768 at 16:9, 1216 x 832 at 3:2 and 896 x 1152 at 3:4.
How to use Hunyuan Image 3 on Pixel Dojo
- 1
Start from a prompt
Type in the box on this page or tap Use this prompt on any example. Lead with the kind of shot, then the place, the people and the light.
- 2
Add every detail you want to see
Hunyuan Image 3 takes prompts up to 3,000 characters, so list the props, wardrobe, weather and textures one by one.
- 3
Pick the shape and the mode
Choose one of seven aspect ratios, then Fast or Quality under Mode. Both cost 2 credits.
- 4
Generate and refine
Compare a few results, keep the one you like and change one detail at a time. Through the API you can also set a seed to repeat a result.
Hunyuan Image 3 pricing: 2 credits per image
Hunyuan Image 3 is included in every Pixel Dojo plan, from $10/month. Every image costs 2 credits from your plan, in Fast or Quality mode and at every aspect ratio.
We made all the Hunyuan images on this page through the Pixel Dojo API at that price, and the same prompts cost the same in the web tool.
Hunyuan Image 3 questions
What is Hunyuan Image 3?
Hunyuan Image 3, also called HunyuanImage 3.0, is Tencent's text to image model, an 80 billion parameter mixture of experts. It is known for photoreal detail and for following long, complex prompts. On Pixel Dojo you can use it in the web tool or through the API for 2 credits per image.
What is Hunyuan Image 3 good at?
Long, detailed prompts with many people and props, photoreal street scenes and portraits, food, architecture and painterly concept art. Our 926 character bookshop cafe prompt came back with all twelve details in place.
How much does Hunyuan Image 3 cost?
Hunyuan Image 3 is 2 credits per image, in Fast or Quality mode and at every aspect ratio. It is included in every Pixel Dojo plan, from $10/month.
What is the difference between Fast and Quality mode?
Fast is the default. Quality spends more steps on sharp edges, fine texture and clean gradients. Both cost 2 credits. Five of the six showcase images on this page are Quality mode, and the cliff library is Fast mode.
Can Hunyuan Image 3 put text in images?
Yes, for short lines. Put the words in double quotes and add "the text must be exact". Our ALPINE EXPRESS travel poster and the SOUP OF THE DAY: PUMPKIN chalkboard printed exactly as written.
What sizes and aspect ratios does Hunyuan Image 3 make?
Seven aspect ratios: 1:1, 16:9, 9:16, 4:3, 3:4, 3:2 and 2:3. Our files measured 1024 x 1024 square, 1280 x 768 at 16:9, 1216 x 832 at 3:2, 1152 x 896 at 4:3, 896 x 1152 at 3:4 and 832 x 1216 at 2:3.
Can I use Hunyuan Image 3 through an API?
Yes. POST a prompt with aspect_ratio, and optionally go_fast and seed, to /api/v1/models/hunyuan-image-3/run. Every image on this page was made through the Pixel Dojo API.
Make your first Hunyuan Image 3 image
Write a prompt above or start from any example on this page. One subscription covers every model on Pixel Dojo.
Start with Hunyuan Image 3