<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/34925661-e0c6-4c63-8180-d6a920efcf94/original=true/34925661-e0c6-4c63-8180-d6a920efcf94.jpeg" />You can get this model through Civitai Early Access, or grab it along with many others by joining The Tinkerer on Whop. Membership gets you early releases, private tools and members-only pages. There are free pages too, no membership needed.👉 <a target="_blank" rel="ugc" href="https://whop.com/cyberdelia-ai-lab">Join on Whop</a> 💬 <a target="_blank" rel="ugc" href="https://discord.gg/88R2hdjb22">Join the community for support, free tools and early news on Discord</a><hr />CyberRealistic Z-Image Turbo is a realism-focused finetune of Z-Image Turbo by Tongyi-MAI.The idea behind it is deliberately simple: keep what makes Z-Image Turbo good - speed, strong prompt understanding, good composition and extremely efficient few-step generation - while moving the default visual language further toward believable photography.CyberRealistic doesn't try to turn Z-Image Turbo into a completely different model. The original already has a very capable photographic foundation. The finetune mainly changes what the model considers a "normal" photograph: more natural skin, less synthetic rendering, more believable faces, stronger material texture, more grounded lighting and better anatomical consistency.Z-Image Turbo already knows how to make a good image. CyberRealistic mainly changes where it starts.<h3 id="what's-different-from-base-z-image-turbo">What's different from base Z-Image Turbo</h3><ul><li>Stronger photographic look out of the box.</li><li>More natural skin texture with less waxy or overly polished rendering.</li><li>Improved faces, eyes, hair and small facial details.</li><li>Better anatomical consistency, especially hands, feet and complex poses.</li><li>Fewer duplicated limbs and extra hands in more difficult compositions.</li><li>More believable fabric, hair, skin, metal, glass and other material textures.</li><li>Stronger response to available light, practical lighting and real-world camera language.</li><li>Less dependence on stacks of words like <code>masterpiece</code>, <code>8k</code>, <code>ultra detailed</code> and <code>photorealistic</code>.</li><li>Keeps the speed and general prompt behavior that make Z-Image Turbo useful.</li></ul>The focus is photography, but that doesn't mean the model is locked to photography. Illustration, cinematic stylization, fantasy, advertising, vintage photography and other looks are still available when you describe them.<hr /><h2 id="prompting">Prompting</h2>If you're coming from SDXL, Pony or Illustrious, the biggest change is simple:Describe the image instead of building a tag stack.Z-Image Turbo uses a Qwen3-based text encoder and responds very well to normal descriptive language. Short comma-separated clauses are completely fine, but every part of the prompt should ideally tell the model something visual.Instead of:<pre><code>woman, realistic, masterpiece, best quality, detailed skin, cinematic, 8k
</code></pre>try:<pre><code>A woman standing beside an open apartment window on a warm summer evening, photographed with soft natural light falling across her face, loose dark hair, natural skin texture and an out-of-focus city street behind her.
</code></pre>The second prompt gives the model an actual scene to construct.<h3 id="put-the-subject-first">Put the subject first</h3>Start with what the image is about.<pre><code>A middle-aged mechanic leaning over the open engine bay of an old red pickup truck...
</code></pre>works better than hiding the subject halfway through a long list of style instructions.You don't need to obsess over exact prompt order, but the main subject and composition should be clear early.<h3 id="be-specific">Be specific</h3>Specific visual language usually does more than generic quality words.Instead of:<pre><code>beautiful lighting
</code></pre>try:<pre><code>soft late-afternoon sunlight entering through a dusty workshop window
</code></pre>Instead of:<pre><code>detailed clothing
</code></pre>try:<pre><code>a faded blue denim jacket with worn seams and slightly frayed cuffs
</code></pre>Instead of:<pre><code>cinematic portrait
</code></pre>try:<pre><code>photographed from chest height with a 50mm lens, shallow depth of field and soft window light from camera left
</code></pre><h3 id="describe-the-light">Describe the light</h3>Lighting is one of the easiest ways to change the realism and mood of the image.Useful examples:<pre><code>soft overcast daylight
</code></pre><pre><code>direct midday sunlight creating hard shadows
</code></pre><pre><code>a single warm tungsten lamp above the table
</code></pre><pre><code>cold fluorescent supermarket lighting
</code></pre><pre><code>late-afternoon sunlight entering through venetian blinds
</code></pre><pre><code>direct on-camera flash in a dark room
</code></pre>You can still use words like <code>cinematic</code>, but describing where the light actually comes from gives the model much more information.<h3 id="quality-tags-are-not-magic-switches">Quality tags are not magic switches</h3>Words such as:<pre><code>masterpiece
best quality
8k
ultra detailed
absurdres
score_9
</code></pre>can still influence the wording of the prompt, but Z-Image Turbo doesn't treat them like the traditional SDXL/Pony quality system.Use that prompt space to describe what you actually want to see.<h3 id="camera-language-works-well">Camera language works well</h3>For photographic images, camera terminology can be useful when it describes a visible effect:<pre><code>35mm documentary photograph
</code></pre><pre><code>85mm portrait lens with shallow depth of field
</code></pre><pre><code>handheld photograph with slight motion blur
</code></pre><pre><code>direct flash snapshot
</code></pre><pre><code>wide-angle environmental portrait
</code></pre><pre><code>medium-format color photograph
</code></pre>Don't feel forced to specify a camera and lens in every prompt. Sometimes simply saying <code>casual phone photo</code> gives you exactly the look you need.<h3 id="prompt-length">Prompt length</h3>There is no perfect prompt length, but these are useful practical ranges:<ul><li>10–30 words: exploration and seed hunting.</li><li>30–80 words: good balance between control and freedom.</li><li>80–150 words: complex scenes, precise lighting or detailed compositions.</li></ul>Long prompts aren't automatically better. Contradictory prompts are the bigger problem.If you ask for soft natural window light, hard direct flash, deep cinematic shadows and flat commercial studio lighting at the same time, the model still has to decide which instruction wins.<h3 id="text-inside-images">Text inside images</h3>Z-Image Turbo is unusually capable at rendering text compared with older diffusion models.If exact text matters, put it in quotation marks:<pre><code>A small neon sign above the diner entrance reading "OPEN ALL NIGHT"
</code></pre>Keep important text reasonably short. It's good, but it still isn't a replacement for a typography application.<hr /><h2 id="recommended-settings">Recommended settings</h2>Z-Image Turbo is a distilled few-step model.Don't treat it like an SDXL checkpoint that needs 30–50 steps.A good starting point is:<ul><li>Steps: 8–9</li><li>CFG / Guidance: effectively OFF</li><li>Resolution: start around 1 megapixel and increase if your hardware allows it</li><li>Negative prompt: normally unnecessary</li></ul>In the original Diffusers implementation, guidance is <code>0.0</code>.In standard ComfyUI workflows, the equivalent no-CFG setup is generally CFG 1.0.More steps are not automatically better with Turbo. If something isn't working, changing the prompt, seed, sampler or composition usually makes more sense than simply increasing the step count.<h3 id="comfyui">ComfyUI</h3>For ComfyUI, I recommend starting with the current Z-Image Turbo workflow/template rather than applying old SDXL settings.Z-Image Turbo has its own sampling behavior and is designed around very low step counts.Negative conditioning is normally zeroed out in the standard Turbo workflow because the model runs without traditional classifier-free guidance.<hr /><h2 id="example-prompts">Example prompts</h2><h3 id="natural-light-portrait">Natural-light portrait</h3><pre><code>A woman in her early thirties sitting beside an open café window, loose brown hair falling across one side of her face, wearing a simple cream-colored sweater. Photographed from slightly below eye level with a 50mm lens, soft overcast daylight entering from the window, natural skin texture, muted colors and a busy street softly blurred in the background.
</code></pre><h3 id="documentary-photography">Documentary photography</h3><pre><code>An elderly fishmonger arranging silver mackerel on crushed ice at an indoor market early in the morning. Cold daylight enters through the open market doors and mixes with the warm bulbs above the counter. Wet concrete floor, weathered hands, faded rubber apron, handheld 35mm documentary photograph with subtle grain and natural color.
</code></pre><h3 id="low-light-snapshot">Low-light snapshot</h3><pre><code>A young woman standing alone beside a vending machine outside a convenience store at two in the morning, photographed with direct on-camera flash. Dark parking lot behind her, slightly messy hair, casual oversized jacket, realistic skin texture, hard flash shadows, muted colors and the imperfect look of a spontaneous late-night photograph.
</code></pre><hr /><h2 id="a-few-last-things">A few last things</h2>Short prompts are completely valid.One of the advantages of Turbo is that you can generate several directions quickly, choose the seed or composition you like, and then add more camera, lighting and material detail.That often works better than trying to write the perfect 150-word prompt before generating anything.Also keep in mind that Z-Image Turbo is distilled for speed. Part of that tradeoff is lower variation than a large non-distilled foundation model. If you keep seeing the same interpretation, change the wording more substantially rather than adding another five quality tags.CyberRealistic Z-Image Turbo is released for people who enjoy generating, experimenting, benchmarking and finding the edges of a model.Feedback is especially useful for difficult poses, multiple people, hands and feet, unusual lighting, text rendering and prompts where the model behaves differently from the original Z-Image Turbo.If you find something interesting — good or bad — let me know.<h3 id="credits">Credits</h3>CyberRealistic Z-Image Turbo is based on Z-Image Turbo by Tongyi-MAI.Z-Image Turbo is released under the Apache 2.0 License. Please follow the applicable upstream license when using or redistributing derived models.

CyberRealistic Z-Image Turbo

Join The Tinkerer on Whop. Membership gets you early releases, private tools and a bunch of extra stuff. 👉 <a target="_blank" rel="ugc" href="https://whop.com/joined/cyberdelia-ai-lab/products/the-tinkerer/">Join on Whop</a>Yes, it jumps from v1.0 to v1.3. No, it’s not a revolution. Just a cleaner, tighter, slightly more caffeinated version of the original. Same soul. Fewer hiccups. More “oh damn, that looks good.”

cyberrealisticZImage_v13_pruned_fp8.safetensors

cyberrealisticZImage_v13_full_bf16.safetensors

v1.3

v9.0

v8.0

v7.0 INT8

v7.0

v6.0

v5.0

v4.0

v3.0

v2.0

v1.0

v0.5

<a target="_blank" rel="ugc" href="https://www.patreon.com/cyberdelia"><img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/3b2f7c16-629e-4545-9ec7-3be16437e462/width=525/3b2f7c16-629e-4545-9ec7-3be16437e462.jpeg" /></a>Join the CyberDelia AI Lab Discord! Hang out with other AI creators, share ideas and resources, get early access to experimental tools and checkpoints, and help shape the growing community. 🚀🔗 <a target="_blank" rel="ugc" href="https://discord.gg/GUByyMuua3">Click here to join</a><hr />CyberRealistic Z-Image Turbo is an early experimental build based on the original Z-Image Turbo.This release should be seen as a technical experiment, not a finished or fully tuned model. It follows the original Z-Image Turbo settings and philosophy, with minimal intervention, to explore how well this setup translates into the CyberRealistic workflow. Big thanks to <a target="_blank" rel="ugc" href="https://huggingface.co/ostris">Ostris </a>for the amazing work - without him, we wouldn’t have been able to do much with this yet.<h3 id="what-this-is-not">What this is not</h3>❌ Not a polished CyberRealistic mainline release ❌ Don’t expect miracles ❌ It won’t turn weak ideas into masterpieces ❌ It won’t solve world hunger ❌ It won’t bring world peace ❌ It will not end wars, climate change, or Monday mornings ❌ It won’t make coffee or do your taxes ❌ It is not sentient, aware, or secretly judging your prompts<h3 id="so...-what-is-this-then">So… what is this then?</h3>✅ A fun experimental build ✅ Fast, responsive, and interesting to play with ✅ Meant for experimentation, comparison, and curiosity ✅ Shared to learn from real-world usage and feedback⚙️ Personal Settings (Forge Neo)<pre><code>Sampler: DPM++ 2s a RF
Schedule: Beta
Sampling steps: 14 Steps
CFG: 1
Shift: 6</code></pre>Essential files: VAE: <a target="_blank" rel="ugc" href="https://huggingface.co/Comfy-Org/z_image_turbo/tree/main/split_files/vae">https://huggingface.co/Comfy-Org/z_image_turbo/tree/main/split_files/vae</a> Text-encoder: <a target="_blank" rel="ugc" href="https://huggingface.co/Comfy-Org/z_image_turbo/tree/main/split_files/text_encoders">https://huggingface.co/Comfy-Org/z_image_turbo/tree/main/split_files/text_encoders</a><hr />This model is shared as-is, for users who enjoy experimenting, benchmarking, and pushing models outside the comfort zone.Feedback, findings, and edge cases are welcome - this release exists primarily to learn from real-world usage.

Yes, it jumps from v1.0 to v1.3.
No, it’s not a revolution.
Just a cleaner, tighter, slightly more caffeinated version of the original.
Same soul. Fewer hiccups. More “oh damn, that looks good.”

CyberRealistic_z_image_v1.3.safetensors

V1.3

Files with SHA256: 77c1b152f049b1b24a2f65e747eb9e74cf91c77cd6d725b74e02c500498b8988