A 20-year-old girl with a sweet baby face but a cool expression, sitting in the passenger seat of a car at night. Wearing a fitted black tube top with a silver heart pendant necklace, and low-rise denim jeans. Skin glistening under direct flash photography, high contrast shadows. Looking at the camera with a playful wink, holding a colorful lollipop. Background is dark car interior with blurred city street lights outside the window. Raw aesthetic, film grain style, Kodak Portra 400 vibe, ultra-realistic skin texture.
sweet yet subtly sexy nighttime aesthetic,silky fabric reflecting soft city lights,bare shoulders highlighted by warm glow,calm and confident expression,elegant curves without exaggeration,clean background with gentle bokeh,social media appropriate sensual vibe
A stylish young woman with silver-dyed hair and a doll-like cute face, wearing a metallic silver puffer vest and a fitted electric purple crop top, over-ear retro headphones, glossy lips, cool but flirtatious gaze. Leaning against a glass railing at night, neon city lights reflecting on her face. Cyberpunk vibe but soft and approachable, shiny fabric textures. One vibrant purple accent, moody cinematic lighting, bokeh city background. Photorealistic, 8K, captured on Sony A7R IV.
A beautiful girl with messy bun hair and soft flushed cheeks, wearing an oversized cream-colored off-shoulder knit sweater, subtly revealing one bare shoulder and collarbone, wearing tiny satin shorts (barely visible), sitting casually on a plush rug. Soft morning sunlight streaming through sheer curtains, warm atmosphere. Gentle seductive gaze, looking directly at camera, relaxed posture. Neutral tones with warm light, lifestyle photography aesthetic. Canon R5, 50mm f/1.2, natural pores and skin texture.
An attractive 20 years chinese girl with long wavy dark hair and clear bright eyes, wearing a reworked school uniform style: a cropped white shirt with a navy blue tie, and a low-waisted plaid mini skirt. Standing in a vintage library aisle, holding a book but looking back over her shoulder. Playful smirk, intellectual but alluring, "good girl gone bad" vibe. Soft indoor library lighting, dust motes dancing in light rays. One deep navy blue accent, rich textures, sharp focus on eyes. Professional portrait photography, 85mm lens.
a gorgeous 20-year-old Western woman,cute innocent face, clear bright eyes,fair smooth skin glowing under city lights,wearing a royal purple satin mini dress,soft fabric gently outlining her figure,standing under evening street lights,body slightly turned, looking back at the camera,cinematic urban background,sweet expression with a hint of allure
a beautiful 20-year-old Western woman,cute youthful face, soft facial features,fair porcelain skin, smooth and glowing,big clear eyes, gentle expression,wearing a high-saturation hot pink fitted top,slim straps, subtly flattering the figure,standing on a lively European street,hands lightly touching her collarbone,urban background with shops and pedestrians,natural daylight, shallow depth of field,sweet, charming, and softly sexy
a beautiful 20-year-old Western girl,cute cheerful face, youthful smile,fair radiant skin,wearing a high-saturation yellow fitted top,light skirt flowing naturally,walking forward mid-step,hands relaxed, playful movement,sunlit park or travel street background,bright, airy, social media aesthetic,sweet, lively, and attractive
adult woman, sweet yet confident feminine energy,wearing a burnt orange fitted crop top with tasteful cut-out detail,showing a hint of waist, elegant and restrained,hands holding a small bag or adjusting accessories,standing in a colorful market street with stalls and signs behind her,soft smile, relaxed shoulders,background rich with daily life details,natural light, social-media-ready composition
cute but alluring style,sweet facial features with a subtle flirtatious gaze,slightly parted glossy lips,sparkling eyes,emotional softness,short fitted top emphasizing natural curves,playful but elegant accessories,gentle movement,candid moment,high-end yet approachable vibe
A stunning young woman browsing in a bright supermarket aisle, cute innocent face with minimal makeup, messy bun hairstyle. Wearing a tight cobalt blue racerback tank top that highlights natural curves, paired with baggy gray sweatpants (contrast of tight & loose). Reaching for a product on the top shelf, body stretched elegantly, mid-movement capture, candid lifestyle photography. Bright fluorescent overhead lighting, colorful shelves in background (blurred). Sony A7IV, 35mm lens, sharp focus on eyes, vibrant colors.
a 20-year-old Western woman, classic European facial features, confident expression, wearing a burnt orange fitted crop top with tasteful cut-out detail, high-waisted light trousers, leaning casually against a café window, one arm raised touching her hair, warm street background with plants, tables, reflections, soft daylight, rich lifestyle atmosphere, sweet with a hint of spice, elegant and approachable
cute but alluring style,sweet facial features with a subtle flirtatious gaze,slightly parted glossy lips,sparkling eyes,emotional softness,short fitted top emphasizing natural curves,playful but elegant accessories,gentle movement,candid moment,high-end yet approachable vibe
A magazine cover of a cheerful 20-year-old Chinese woman with messy long hair tied with a silk ribbon, laughing while sitting at a vintage wooden table in a seaside cafe. She wears a vintage floral sundress and retro sunglasses on her head. An iced coffee and a half-eaten croissant are on the table. Dappled sunlight filtering through leaves, warm golden hour glow, Kodak Gold 200 film aesthetic, slight halation, nostalgic mood, candid snapshot style. 8K resolution. Magazine layout: Title "SUNDAY". Cover text: "Golden Hour", "Slow Living", "July Edition 2025". Barcode bottom. Playful retro 70s typography in orange and cream.
Create a premium 3D rendered Garfield collectible vinyl figure in a modern designer toy aesthetic. The figure should feature glossy translucent vinyl material with subtle light refraction, bold saturated colors true to the character, and simplified geometric forms with smooth curves. Position against a pristine gradient background transitioning from light gray to white. Use professional product photography lighting with soft key light from above and gentle rim lighting to highlight the vinyl’s glossy finish. The character should be posed in a confident standing position, centered perfectly in frame. Apply shallow depth of field with the figure in sharp focus. Render in ultra-high resolution with clean minimalist composition, no text, logos, or distracting elements. Square aspect ratio 1080x1080 pixels, photorealistic quality with crisp details and vibrant color reproduction suitable for premium toy marketing.
A 9:16 vertical screen high-end fashion illustration mood board, simulating a tablet scan effect. The background is pure hand-drawn creamy watercolor gradient paper with a faint pink grid. The visual core consists of several glossy vinyl stickers with distinct white die-cut wide borders and soft shadows. The central sticker is a photo of the user wearing a sweet date outfit, with bright lighting. On the left side is a deconstructed sticker of this outfit: a neatly folded jacket and exquisite high heels. In the bottom right corner is the key hidden layer sticker: a chic open mini-handbag revealing daily essentials like a tube of lipstick and vintage sunglasses, showcasing leather and glass textures. A Labubu art doll sticker in pink tones that echoes the user's clothing is lying on a hand-drawn speech bubble. The surroundings are decorated with crayon-textured hand-drawn hearts, sparkle symbols, and scribbled Chinese calligraphy annotations for OOTD. The image contains absolutely no human hands, pens, or physical desktop backgrounds—pure flat art illustration.
A lithograph poster of a majestic elk standing in a dense redwood forest, printed in vintage Forest Green and Burnt Orange inks with posterized shading. Includes stylized text "WILDERNESS", worn corners, and faded paper texture like a 1960s print ad.
1girl, a girl as a lifeguard, sitting in the sun on a white lifeguard tower, plain white shirt, red shorts, very bright and sunny, watching over the beach, natural hair color and eyes, 8k, professionally color graded, depth of field
Chinese mythology character combination illustration, featuring the three classic characters Erlang Shen, Sun Wukong, and Nezha, in a Q-version cute style, dynamic and lively. - Erlang Shen: Calm and composed expression, with the third eye on his forehead slightly closed, wearing an ornate golden crown, dressed in exquisite traditional battle robes, holding a three-pronged, two-edged sword, accompanied by a cheerful and adorably dazed Xiaotian Dog at his side. - Sun Wukong: Confident and mischievous expression, wearing a phoenix-winged purple gold crown on his head (typically a hair-binding crown with pheasant feathers, resembling two "cockroach whiskers"), with fluffy and stylish golden monkey fur, dressed in a yellow tiger-skin short skirt and auspicious cloud battle armor, gripping the Ruyi Jingu Bang, striking the classic pose of gazing into the distance, spirited and proud. - Nezha: Playful and brave facial expression, with two sky-high pigtails, wearing flowing and ethereal red lotus flower battle armor, standing on wind-and-fire wheels, holding a fire-tipped spear, with the Universe Ring encircling his body, full of the aura of a young hero. Overall painting style is delicate and refined, with soft warm color tones, clear and fluid lines, carrying a subtle watercolor illustration texture, simple and elegant background, the scene filled with fun, warmth, and storytelling.
A photo taken with an extreme fisheye lens. A young woman with blonde twin tails wearing a gray cardigan and plaid skirt school uniform is excitedly jumping at the Shibuya Scramble Crossing, with one hand dramatically reaching toward the foreground of the lens, her fingernails clearly visible. In the background, the distorted Shibuya 109 building and other structures stand tall, the streets crowded with pedestrians and vehicles. A huge pink and blue gradient cartoon monster floats above the city, with massive tentacles and horns surrounding the distorted cityscape. Sunny weather with strong light and shadow contrasts. Circular frame.
From rapid prototyping with Turbo to final delivery with Base. Discover how dual-model flexibility transforms industries
Create studio-quality assets in seconds.
Generate realistic product backdrops and marketing materials instantly. The Z-Image model's ability to follow complex prompts ensures your product is showcased in the exact lighting and environment you envision, drastically reducing photoshoot costs.
Visuals that understand your culture.
Leverage the best-in-class Chinese language support to create Wuxia styles, Hanfu portraits, or culturally specific artwork. Z-Image interprets nuanced cultural descriptors that other models simply ignore.
Iterate faster than ever before.
With 8-step generation, concept artists can generate dozens of variations per minute. Whether you need character sheets, environmental textures, or UI elements, Z-Image fits perfectly into a fast-paced agile development pipeline.
Design with embedded text support.
Create posters, book covers, and social media banners where text and image blend seamlessly. Z-Image's superior OCR-like understanding of text generation helps designers overcome one of AI's biggest hurdles.
Select between creative flexibility and lightning-fast production
Foundation Model for Creative Freedom
The undistilled foundation model designed for maximum creative control. Supports full Classifier-Free Guidance (CFG), negative prompting, and delivers high diversity across seeds. Ideal for professional workflows requiring precise prompt engineering and fine-tuning.
Best For
Distilled for Speed & Quality
A distilled variant optimized for rapid generation. Achieves very high visual quality in just 8 steps through Reinforcement Learning optimization. Perfect for rapid prototyping and production workflows where speed is critical.
Best For
| Capability | Z-Image (Foundation) | Z-Image Turbo |
|---|---|---|
| Model Architecture | S3-DiT | S3-DiT |
| Distillation | Undistilled (Full) | Distilled (8-step) |
| Inference Steps | 28–50 steps | 8 steps only |
| Classifier-Free Guidance (CFG) | ✅ Full support | ❌ Not available |
| Negative Prompting | ✅ Supported | ❌ Not supported |
| Output Diversity | High (varied seeds) | Low (consistent) |
| Visual Quality (RL) | High | Very High (RL optimized) |
| Fine-tuning (LoRA/ControlNet) | ✅ Ready for training | ❌ Not recommended |
| Best For | Complex workflows | Speed & consistency |
Use Turbo for rapid prototyping (8 steps), then switch to Z-Image for final refinement with CFG and negative prompts.
Try Z-Image On ArenaFrom lightning-fast inference to deep customization. Discover the perfect version of Z-Image for your workflow.
Lightning-fast text-to-image generation with enterprise-grade efficiency
Ultra-fast inference
Parameters
Efficient memory
Creative AI · Free trial · Instant generation
The Complete Z-Image Ecosystem: Z-Image Base delivers the full 6B parameter power for production-grade photorealism, while Z-Image Turbo brings distilled 8-step efficiency for rapid iteration. Both share Tongyi-MAI's bilingual DNA and consumer-grade hardware optimization.
The foundational 6B parameter model delivers high-detail generation with superior texture rendering and complex lighting simulation. Ideal for final assets and professional workflows where photorealistic precision is paramount.
Advanced adversarial distillation reduces generation to 8 steps, delivering 95% of Base quality in sub-second latency. Perfect for real-time applications, rapid prototyping, and high-volume creative pipelines without sacrificing core visual fidelity.
Trained on massive Chinese and English corpora, both models understand cultural context, historical references, and idioms natively—not just translation. Accurately interprets complex Eastern and Western literary concepts that generic models misinterpret.
Eliminate the 'plastic AI look' with authentic skin textures, natural subsurface scattering, and film-grain aesthetics. Z-Image Base excels at museum-grade portraiture, while Turbo maintains impressive realism at unprecedented speeds.
Efficient 6B architecture runs smoothly on 12GB+ VRAM consumer GPUs (RTX 3060/4060 and up). Significantly lighter than Flux-12B, making professional-grade AI generation accessible without enterprise hardware investments.
Generate images with legible, stylistically accurate embedded text and signage. Perfect for poster design, logo concepts, and marketing materials where visual and textual elements must blend seamlessly—a historical weak point for diffusion models.
Turbo is optimized for speed, generating images in just 8 steps (milliseconds). Base is optimized for maximum quality and detail, taking slightly longer but delivering superior textures and lighting for professional use.
Use **Base** when quality is critical: final marketing assets, detailed illustrations, or complex multi-subject scenes. Use **Turbo** when speed matters: brainstorming, real-time iteration, or generating dozens of variations quickly. You can switch between them instantly on our platform.
Currently, z-image.app hosts both the Base and Turbo models for immediate use. The Z-Image Edit model (for in-painting and image manipulation) is in our technical preview phase and will be available to Pro users soon.
Z-Image Turbo is a next-generation text-to-image model developed by the Tongyi team. Its main differentiator is the 'Turbo' technology (distillation), which allows it to generate high-quality images in just 8 steps. This makes it significantly faster than traditional models like SDXL or Flux while maintaining exceptional image fidelity.
Yes, our online playground is currently free for public use. To ensure fair access and stability for everyone, we use Cloudflare Turnstile to prevent bot abuse. We plan to introduce a subscription tier in the future for power users who need higher limits, API access, or faster queues.
Absolutely. This is one of Z-Image's 'Killer Features'. Unlike Midjourney or Flux which are optimized for English, Z-Image is natively trained on a vast corpus of bilingual data, offering superior understanding of Chinese prompts, idioms, and cultural contexts.
Yes. The Z-Image model weights are released under the Apache 2.0 license, which is highly permissive and allows for commercial use. You own the images you generate on our platform.
Since Z-Image Turbo is extremely fast and popular, we use Cloudflare Turnstile to protect our GPU resources from automated bots and DDoS attacks. This ensures that real human users like you get the fastest possible generation experience.
Z-Image Turbo is designed to be lighter (6B vs 12B parameters) and much faster (8 steps vs 20-50 steps). While Flux is excellent, Z-Image offers a better balance of speed and quality for users with standard hardware, and provides significantly better support for Chinese language inputs.
The Z-Image family includes Turbo, Base, and Edit versions. Currently, our site hosts the Turbo version for maximum speed. We are technically ready to deploy the Base (for fine-tuning enthusiasts) and Edit (for image manipulation) models as soon as they are fully released to the open-source community.
Z-Image is efficient. You can run the Turbo model locally on consumer graphics cards with as little as 12GB of VRAM (e.g., RTX 3060/4060 or higher). This makes it much more accessible than larger models that require 24GB+ of video memory.
From Z-Image to Nano Banana Pro, find the AI image generation tool that fits your needs
Open weights for boundless developer creativity.
Balance speed and quality on demand.
Native 4MP resolution for commercial consistency.
Your Vision, Exactly As Pictured
Reasoning AI for perfect text and physics.
Improved version of Qwen Image
AI editing with perfect consistency
Photo editing AI that decomposes images into layers for precise, non-destructive edits.
Cinematic lighting for dreamlike portraits.
The uncompromising 6B parameter foundation model. Masterpiece-level photorealism and native bilingual understanding.
Real-time photorealism in under a second.
Don't let slow models kill your inspiration. Z-Image Turbo turns your imagination into photorealistic art in milliseconds. No complex setup, just pure creativity.
