Hi! The model is indeed interesting, but the realism still needs some fine-tuning. I really like how it generates images—though the contrast can sometimes be way too high.

But I have a question for you: do you really think most people, after generating an image with a neural network, will rush into Photoshop just to tweak the contrast? To me, that’s just unnecessary hassle. Speaking personally, I create images for fun and visual enjoyment—and I’m not willing to spend extra time editing something in post-production when the same result could be achieved by simply adjusting that parameter in the next model version, as the author has already done.

After all, most of us here are just having fun, creating all sorts of whimsical or surreal images. For serious professional work, I believe experts use different tools altogether. AI models are evolving rapidly, and yes—it’s already becoming hard to tell whether a person in an image is real or generated. But it’s still possible… for now. I reckon in another couple of years, it’ll be extremely difficult to distinguish the two.

@Cyberdelia Hi! Yes, I’ve already noticed that the contrast has been reduced in v6—thanks for that! I’ll give it a try.

@Zakman99 Thank you for the comment!

That's a fair point. I think it really comes down to personal preference and workflow. As I mentioned, this is still one of the best models I've tested so far.

To be fair, I don't think most users will even mind the stronger contrast. In some images, I actually think it looks fantastic and adds a lot of impact. In my case, I'm just very used to a more balanced look, so I often do a quick adjustment in Photoshop, slightly reducing the contrast or lifting the blacks a bit. That usually gets the image exactly where I want it.

So for me it's more of a personal preference than a major issue. The model is already incredible as it is. Hopefully future updates can refine the contrast a little further, because if that gets improved, something that's already excellent could become even better.

And yes, I also agree with your point about realism. AI image generation is improving incredibly fast, and the gap between real and generated images keeps getting smaller every year. There are still some telltale signs today, but for many casual viewers it's already becoming difficult to tell the difference. I wouldn't be surprised if future versions make that distinction even harder.

Thanks again for sharing your thoughts!

@Cyberdelia Thank you for the update!

Yes, I'm currently using V5, and it has quickly become one of my favorite models. The image quality and prompt adherence have been excellent.

I'm especially happy to hear that the contrast has been toned down in V6. That was one of the very few things I found myself adjusting occasionally, so it's great to see that you're continuously refining the model based on community feedback.

Thank you for all your hard work on this model and for taking the time to engage with the community. I'm looking forward to testing V6!

@Cyberdelia danbooru prompts, but probably also training images. Like your cyberrealistic pony model. I guess it's fine-tuned with danbooru tagged images.
Oh, and I would also like to be able to do negative prompting. My standard negative prompts include futa and loli.

@Cyberdelia "anime" did not sound like it is what I'm looking for, and the images look too plastic to me. I do like the more realistic results of this model here, but the prompting is too limited. And without negative prompts, I do get futa results too often.

@Cyberdelia example where it went wrong: https://civitai.red/posts/29265886
prompting the pony model worked much more reliable IMHO.

@Cyberdelia thank you (sent you a tip earlier, all the yellow buzz that I have, as I don't have a paid account). I guess this was using your prompt enhancer workflow. That did indeed improve results substantially. Both with the pony model and this one, there is one thing I still cannot reliably achieve, even with prompt enhancing: deep penetration.

I like that. Can't wait to try it. I think the color is great.

Thank you for sharing your experience with V6!

It's great to hear that the colors, gamma, and overall realism have been improved. The stronger contrast was one of the few things I occasionally adjusted in V5, so your feedback is very encouraging.

I also want to thank the creator for listening to community feedback and continuing to improve the model. It's always great to see a developer actively engaging with users and refining an already excellent model.

I really appreciate you taking the time to test it and share your thoughts with the community. Comments like yours are very helpful for those of us who are still waiting to try V6 ourselves.

Thanks again to both of you!

@Cyberdelia Hey, thanks for you answer and support! I've the link now :)

Agreed. Version 6 is a huge step backward from what version 5 was capable of.

@soyv4 I'm asking out of pure curiosity—what is it that isn't audible? Why is it blurred? I didn't notice anything like that.

@PrincesOfDarkness Try this model and you will understand for yourself. https://civitai.red/models/2237711/z-image-turbo-nsfw-photorealistic-zit?modelVersionId=3060630

@soyv4  ok try . But what specific categories does this checkpoint understand? I mean, what should I test it on—aside from NSFW, of course, since the base model isn't designed for that? For example: fantasy, realistic, photorealistic, artistic... what should I check?

@PrincesOfDarkness The Z-Image Turbo NSFW v9.3 model works with any category, just try it.

Subject: Detailed feedback: Z-Image Turbo NSFW v9.3 vs. CyberRealistic Z-Image Turbo v6.0

Hi! I’ve been testing both models using the exact same prompt (a cyberpunk angel warrior scene). Here is my detailed breakdown of the results:

Z-Image Turbo NSFW v9.3:

Material Rendering: Metal looks like real steel with scratches, dirt, and realistic environmental reflections. It has a very authentic, "lived-in" look.

Lighting & Glow: The lighting is physically correct. It uses warm fire tones, natural shadows, and realistic glow effects that feel grounded in reality.

Mechanical Details: Incredible complexity in the wings and joints. You can clearly see hydraulics, bolts, and intricate machinery without it looking messy.

Face & Skin: Gritty, realistic skin texture. The facial expressions feel intense, alive, and human.

Atmosphere: Cinematic and photorealistic. The fire, smoke, and debris behave naturally, giving it a blockbuster movie feel.

CyberRealistic Z-Image Turbo v6.0:

Material Rendering: The metal looks overly glossy, resembling plastic or polished chrome. It lacks the surface imperfections needed for true realism.

Lighting & Glow: It relies heavily on "acidic" neon blue lighting. It feels artificial, over-saturated, and less grounded in a real environment.

Mechanical Details: The wings and armor look more like a stylized 3D render or concept art. The details are somewhat simplified compared to v9.3.

Face & Skin: The skin feels "doll-like" or smoothed out. The facial expressions are a bit more rigid and less dynamic.

Atmosphere: It feels more like digital concept art or a comic book style rather than true photorealism.

Verdict:

Z-Image Turbo NSFW v9.3 is a massive upgrade. It completely outperforms v6.0 in terms of photorealism, texture fidelity, and lighting. Great work on the new version!

@PrincesOfDarkness Here are 2 promts to check:

A cellphone photo of african adult woman with pink Gilded Waves hair. She is sexy. She is wearing glasses. Photorealistic, ultra-detailed, 35mm f/1.4 lens, 8K resolution, soft diffused natural window lighting with subtle rim light. One female, 24 years old, slender athletic build, East Asian ethnicity. Almond-shaped hazel eyes, long straight black hair loosely falling over shoulders, neutral seductive expression with slightly parted lips. Skin base tone

In a in a Japanese tea garden, a serene, tranquil setting location with detailed background environment.

In a on kitchen counter location with detailed background environment.

Here's more:

Ultra-detailed dynamic portrait of a biomechanical angel leaping from collapsing ruins, her massive steel wings fully extended mid-flight with razor-sharp feathers scattering. Her living armor shifts with hydraulic precision, glowing energy veins pulsing across articulated plates as she reaches forward in battle-ready stance. Beautiful yet fierce face shows focused determination, silver hair whipping across luminous eyes, lips parted in a war cry. Behind her, a cathedral crumbles - shattered pillars, molten cracks in the ground, and burning embers in the ash-filled sky create apocalyptic contrast. Art by Greg Rutkowski and Wei Wang, hyper-realistic 8K Unreal Engine 5 render, photorealistic textures, dramatic volumetric lighting, intricate biomechanical details, masterpiece fantasy concept art.

@soyv4  ok you win i I tried out that prompt, and the "cyber" version really does provide fewer technical details; I'll test that one in other prompts :)
your promt :P
A detailed technical specification depicts an ultra-detailed dynamic portrait of a biomechanical female humanoid angel entity leaping from collapsing ruins, her massive steel wings fully extended mid-flight with razor-sharp feathers scattering. The character establishes a forward-propelling ballistic trajectory that originates from the geometric center of collapsing neo-gothic architectural ruins. Her torso tilts forward at a precise forty-five-degree angle relative to the horizon line while her left leg fully extends backward to provide aerodynamic stabilization. Her right knee is flexed forward at a sharp ninety-degree angle, and her right arm reaches directly toward the camera lens plane with fingers splayed in a tactical grasping gesture. Her left arm is drawn back parallel to her flank to counterbalance the primary kinetic motion of the leap. The facial structure is Caucasian with a sharp angular jawline and high cheekbones, displaying a beautiful yet fierce face that shows focused determination with silver hair whipping across luminous eyes, and lips parted in a war cry. Her eyes feature luminous, self-illuminating silver-blue irises driven by an internal light effect that casts a subtle volumetric glow onto the surrounding orbital skin. Her hair consists of thousands of individually rendered silver-platinum strands showing high-velocity aerodynamic displacement as they whip laterally across her face, utilizing precise transparency layering for realistic light transmission.The entity possesses a complex endoskeleton integrated seamlessly with living organic tissue, covered by an outer carapace of segmented, articulated titanium-alloy plates layered over a carbon-fiber composite under-suit. Her living armor shifts with hydraulic precision, glowing energy veins pulsing across articulated plates as she reaches forward in battle-ready stance. Every armor junction features visible operational mechanical components including micro-hydraulic pistons, exposed pneumatic lines, braided steel coolant hoses, and miniature rotary dampers. The surface of the armor exhibits hyper-detailed industrial wear consisting of laser-etched serial numbers, manufacturing barcodes, superficial combat abrasions, localized paint chipping, and dark grease stains accumulated around mechanical oil seals. An intricate network of translucent subdermal energetic conduits pulses across the articulated chest plate and shoulder pauldrons, cycling dynamically between cyan and emerald green luminescence to indicate shifting power loads in perfect synchronization with the physical movement of the armor.Two massive multi-segmented biomechanical wing structures are attached to the thoracic vertebrae via reinforced mechanical gimbals. The primary structural wing spars consist of dark brushed steel secured by visible structural rivets, hydraulic actuators, and exposed wiring harnesses sheathed in braided Kevlar. The upper wing joints and primary structural spars feature intricate filigree engraving that is directly etched into the dark steel casing via high-precision fiber lasers, exhibiting a complex fusion of gothic cathedral geometry and micro-circuit board traces. Gilded gold-leaf accents precisely outline the mechanical perimeters of the upper wing shoulders to create a stark contrast against the weathered industrial metal. Micro-inscribed runes and technical warning labels reading CAUTION HIGH VOLTAGE and ACTUATOR VENT are stamped in matte white lettering along the margins of the metallic plumage. The edges of the primary feathers are treated with an anodized titanium-nitride coating that creates a subtle iridescent gold-to-violet color shift when catching directional light. The metallic feathers are arranged in three overlapping tiers called primaries, secondaries, and coverts, with each individual feather connected to an independent micro-servo motor allowing autonomous yaw, pitch, and roll adjustments. During the mid-flight leap, exactly forty-two detached razor feathers are captured mid-air, suspended in a state of high-velocity kinetic dispersal behind the character while displaying distinct motion blur and sharp metallic specular highlights.Behind her, a cathedral crumbles as shattered pillars, molten cracks in the ground, and burning embers in the ash-filled sky create an apocalyptic contrast. The surrounding environment consists of a collapsing neo-gothic cathedral interior transitioning into an outdoor apocalyptic wasteland filled with fractured granite pillars, shattered stained glass fragments caught mid-fall, and pulverized concrete debris undergoing rapid downward gravitational acceleration. The ground plane features deep jagged seismic fissures that expose subterranean magma pipes rendered with high-temperature emission properties that create local amber illumination. The atmosphere is a dust-choked and ash-filled troposphere where the sky box displays a dense columnar pyrocumulus cloud formation illuminated from below by raging ground fires. Burning embers of varying sizes drift through the volumetric air space, driven entirely by chaotic thermal updrafts.The final visual composition is delivered as a hyper-realistic 8K Unreal Engine 5 render, featuring photorealistic textures, dramatic volumetric lighting, and intricate biomechanical details to create a masterpiece fantasy concept art piece. The artistic style is heavily influenced by the distinct visual techniques of Greg Rutkowski and Wei Wang, blending classical painterly lighting choices with modern digital precision. The primary light source is an intense low-angle directional key light from the collapsing infrastructure that casts warm high-contrast amber and orange light onto the left side of the character. A secondary light source consists of high-intensity cool cyan fill light emanating from the armor energy veins and distant atmospheric lightning, creating sharp rim-lighting along the right contours of the steel wings. Dense particulate smoke and heterogeneous fog banks with varying density fields allow dramatic volumetric lighting to slice through the shattered cathedral roof structure. The camera positioning creates a shallow depth of field that focuses sharply on the character eyes while blurring the distant background ruins, ensuring an uncompressed high-dynamic-range presentation with rich shadow delineation and exceptional clarity across all primary meshes.

Try constructing the prompt like this and let me know if the result is incorrect.:


A high-fidelity, authentic [CAMERA TYPE & FILM STOCK, e.g., 35mm analog film photograph captured with the deliberate color science, rich tonal range, soft highlight roll-off, and fine organic grain structure characteristic of Kodak Portra 400], portraying an intensely atmospheric and moody [TIME OF DAY & MOOD, e.g., morning scene] inside a [LOCATION STYLE & ENVIRONMENT, e.g., minimalist luxury bedroom].

The camera uses a masterful [SHOT TYPE & CAMERA ANGLE, e.g., medium-wide shot from a low eye-level perspective] to capture the full environment and the entire silhouette of a single subject from head to toe.

The central figure is a stunning [SUBJECT DEMOGRAPHICS & CORE FEATURES, e.g., multi-ethnic woman with voluminous, dark afro-textured curls framing her face], showcasing flawless natural skin with [SKIN TEXTURE DETAILS, e.g., visible pores and a soft satin texture].

She is [EXACT POSITIONING & INTERACTION WITH ENVIRONMENT, e.g., standing directly in front of a large open window, positioned with her back to the glass so her lower body gently rests against the white wooden windowsill].

She is attired in [DETAILED CLOTHING/GARMENT DESCRIPTION, e.g., a fully matching black lace lingerie set: a delicate black lace bra with thin strap details across her chest, a matching high-waisted black lace garter belt with adjustable straps wrapping tightly around her hips, and matching intricate black lace panties that align perfectly with the garter straps]. Her lower limbs are covered only by [SECONDARY CLOTHING/ACCESSORIES DETAILS, e.g., sheer black stockings with wide floral lace tops that attach directly to the garter clips on her thighs].

Her arms are [SPECIFIC BODY POSE & ARMS POSITION, e.g., crossed loosely over her chest in an elegant, defensive yet relaxed posture]. Her head is [HEAD POSITION, FACIAL EXPRESSION & EMOTIONAL INTENSITY, e.g., slightly turned in a three-quarter angle as she directs a deeply focused, calm, and contemplative expression with a subtle, quiet intensity looking off to the side].

The lighting is dominated by a majestic [PRIMARY LIGHT SOURCE & EFFECT, e.g., golden hour morning sun blasting from behind her through the window, creating a powerful, intense rim-light that dramatically traces the contours of her curly hair and shoulders], while the light diffuses beautifully across the [AMBIENT LIGHT SPREAD, e.g., clean minimalist room]. Her front profile and face are softly illuminated by a gentle [SECONDARY/BOUNCED LIGHT SOURCE, e.g., low-contrast wrap of bounced warm light, creating soft subtle shadows].

The expanded background features a sophisticated [BACKGROUND LAYOUT & MAIN FURNITURE, e.g., minimalist bedroom layout with a low-profile bed covered in smooth rumpled white satin sheets], a clean [BACKGROUND DETAILS/PROPS 1, e.g., wooden bedside table holding a small delicate ceramic vase with fresh morning flowers], and a few [BACKGROUND DETAILS/PROPS 2, e.g., structural indoor potted plants with green broad leaves positioned thoughtfully near the corner to catch the ambient glow], resulting in an incredibly rich analog film texture and completely unposed fashion-editorial reality.

@PrincesOfDarkness I've been comparing models using the exact same prompt: Z-Image Turbo NSFW v9.3 and CyberRealistic Z-Image Turbo v6.0. Here are my thoughts:

Z-Image Turbo NSFW v9.3:

• Skin: Visible pores, freckles, natural texture.

• Lighting: High contrast, sunny, sharp shadows.

• Details: Hair and clothing folds are crisp and clear.

• Overall impression: True photorealism, looks like a shot from a good camera.

CyberRealistic Z-Image Turbo v6.0:

• Skin: Smoother, almost "plastic", lacks micro-textures.

• Lighting: Soft, diffused, less dramatic.

• Details: Softer edges, lacks sharpness.

• Overall impression: Looks like a photo with a "beauty" filter applied, or slightly out of focus.

@soyv4  https://civitai.red/models/2625788/realismengine-reprise-40 test this model , And I think you'll like other works by this creator.

,

@PrincesOfDarkness fp8 ? no thanks

@soyv4  https://civitai.red/models/2625788/realismengine-reprise-40?modelVersionId=2962504 normal or https://civitai.red/models/2730247/astrasupreme-real

@PrincesOfDarkness No thanks, I'm happy with what suits me, and there's no need to download what I don't need.