[HN Gopher] Qwen-Image-2.0: Professional infographics, exquisite...
       ___________________________________________________________________
        
       Qwen-Image-2.0: Professional infographics, exquisite photorealism
        
       Author : meetpateltech
       Score  : 359 points
       Date   : 2026-02-10 09:19 UTC (13 hours ago)
        
 (HTM) web link (qwen.ai)
 (TXT) w3m dump (qwen.ai)
        
       | Deukhoofd wrote:
       | The text rendering is quite impressive, but is it just me or do
       | all these generated 'realistic' images have a distinctly uncanny
       | feel to it. I can't quite put my finger on it what it is, but
       | they just feel off to me.
        
         | belter wrote:
         | Agree, looks like the same effect they are applying on YouTube
         | Shorts...
        
         | BoredPositron wrote:
         | Qwen always suffered from their subpar rope implementation and
         | qwen 2 seems to suffer from it as well. The uncanny feel is
         | down to the sparsity of text to image token and the higher in
         | resolution you go the worse it gets. It's why you can't take
         | the higher ends of the MP numbers serious no matter the model.
         | At the moment there is no model that can go for 4k without
         | problems you will always get high frequency artifacts.
        
         | likium wrote:
         | At least for the real life pictures, there's no depth of field.
         | Everything is crystal clear like it's composited.
        
           | derefr wrote:
           | > like its composited
           | 
           | Like focus stacking, specifically.
           | 
           | I'm always surprised when people bother to point out more-
           | subtle flaws in AI images as "tells", when the "depth-of-
           | field problem" is so easily spotted, and has been there in
           | every AI image ever since the earliest models.
        
             | Mashimo wrote:
             | I had no problems getting images with blurry background
             | with the appropriate prompts. Something like "shallow depth
             | of fields, bokeh, DSLR" can lead to good results. https://c
             | dn.discordapp.com/attachments/1180506623475720222/1... [0]
             | 
             | But I found that that results in more professional looking
             | images, and not more realistic photos.
             | 
             | Adding something like "selfy, Instagram, low resolution,
             | flash" can lead to a .. worse image that looks more
             | realistic.
             | 
             | [0] I think I did this one with z image turbo on my 4060 ti
        
               | afro88 wrote:
               | The blur isn't correct though. Like the amount of blur is
               | wrong for the distance, zoom amount etc. So the depth of
               | field is really wrong even if it conforms to "subject
               | crisp, background blurred"
        
               | derefr wrote:
               | Exactly.
               | 
               | My personal mechanistic understanding of diffusion models
               | is that, "under the hood", the core thing they're doing,
               | at every step and in every layer, is a kind of
               | _apophenia_ -- i.e. they recognize patterns /textures
               | they "know" within noise, and then they nudge the noise
               | (least-recognizable pixels) in the image toward the
               | closest of those learned patterns/textures, "snapping"
               | those pixels into high-activation parts of their trained-
               | in texture-space (with any text-prompt input just adding
               | a probabilistic bias toward recognizing/interpreting the
               | noise in certain parts of the image as belonging to
               | certain patterns/textures.)
               | 
               | I like to think of these patterns/textures that diffusion
               | models learn as "brush presets", in the Photoshop sense
               | of the term: a "brush" (i.e. a specific texture or
               | pattern), but locked into a specific size, roughness,
               | intensity, rotation angle, etc.
               | 
               | Due to the way training backpropagation works (and
               | presuming a large-enough training dataset), each of these
               | "brush presets" that a diffusion model learns, will
               | always end up learned as a kind of "archetype" of that
               | brush preset. Out of a collection of examples in the
               | training data where uses of that "brush preset" appear
               | with varying degrees of slightly-wrong-size, slightly-
               | wrong-intensity, slightly-out-of-focus-ness, etc, the
               | model is inevitably going to learn most from the "central
               | examples" in that example cluster, and distill away any
               | parts of the example cluster that are less shared. So
               | whenever a diffusion model recognizes a given one of its
               | known brush presets in an image and snaps pixels toward
               | it, the direction it's moving those pixels will always be
               | toward that archetypal distilled version of that brush
               | preset: the resultant texture in perfect focus, and at a
               | very specific size, intensity, etc.
               | 
               | This also means that diffusion models learn brushes at
               | distinctively-different scales / rotation angles / etc as
               | entirely distinct brush presets. Diffusion models have no
               | way to recognize/repair toward "a size-resampled copy of"
               | one of their learned brush presets. And due to this,
               | diffusion models will never learn to render in details
               | small enough that the high-frequency components of of
               | their recognizable textural-detail would be lost below
               | the Nyquist floor (which is why they suck so much at
               | drawing crowds, tiny letters on signs, etc.) And they
               | will also never learn to recognize or reproduce visual
               | distortions like moire or ringing, that occur when things
               | get rescaled to the point that beat-frequencies appear in
               | their high-frequency components.
               | 
               | Which means that:
               | 
               | - When you instruct a diffusion model that an image
               | should have "low depth-of-field", what you're really
               | telling it is that it should use a "smooth-blur brush
               | preset" to paint in the background details.
               | 
               | - And even if you ask for depth-of-field, everything in
               | what a diffusion model thinks of as the "foreground" of
               | an image will always have this surreal perfect focus,
               | where all the textures are perfectly evident.
               | 
               | - ...and that'll be true, even when it doesn't make sense
               | for the textures to be evident at all, because in real
               | life, at the distance the subject is from the "camera" in
               | the image, the presumed textures would actually be so
               | small as to be lost below the Nyquist floor at anything
               | other than a macro-zoom scale.
               | 
               | These last two problems combine to create an effect
               | that's totally unlike real photography, but is actually
               | (unintentionally) quite similar to how digital artists
               | tend to texture video-game characters for "tactile
               | legibility." Just like how you can clearly see the crisp
               | texture of e.g. denim on Mario's overalls (because the
               | artist wanted to make it feel like you're looking at
               | denim, even though you _shouldn 't_ be able to see those
               | kinds of details at the scaling and distance Mario is
               | from the camera), diffusion models will paint anything
               | described as "jeans" or "denim" as having a crisply-
               | evident denim texture, despite that being the totally
               | wrong scale.
               | 
               | It's effectively a "doll clothes" effect -- i.e. what you
               | get when you take materials used to make full-scale
               | clothing, cut tiny scraps of those materials to make a
               | much smaller version of that clothing, put them on a
               | doll, and then take pictures far closer to the doll, such
               | that the clothing's _material textural detail_ is visibly
               | far larger relative to the  "model" than it should be.
               | Except, instead of just applying to the clothing, it
               | applies to _every_ texture in the scene. You can see the
               | pores on a person 's face, and the individual hairs on
               | their head, despite the person standing five feet away
               | from the camera. Nothing is ever aliased down into a
               | visual aggregate texture -- _until_ a subject gets
               | distant _enough_ that the recognition _maybe_ snaps over
               | to using entirely different  "brush preset" learned
               | specifically on visual aggregate textures.
        
           | albumen wrote:
           | Every photoreal image on the demo page has depth of field,
           | it's just subtle.
        
           | vunderba wrote:
           | Which is pretty amusing - because it's the exact opposite
           | problem that BFL had with the original Flux model - every
           | single image looked like it was taken with a 200mm f/4.
        
         | GaggiX wrote:
         | For me the only model that can really generate realistic images
         | is nano banana pro (also known as gemini-3-pro-image). Other
         | models are closing the gap, this one is pretty meh in my
         | opinion in realistic images.
        
           | Mashimo wrote:
           | You can get flux and maybe z-image to do so, but you have to
           | experiment with the promt a bit. Or maybe get an LoRa to
           | help.
        
             | cubefox wrote:
             | The examples I saw of z-image look much more realistic than
             | Nano Banana Pro, which is likely using Imagen 4 (plus
             | editing) internally, which isn't very realistic. But Nano
             | Banana Pro has obviously much better prompt alignment than
             | something like z-image.
        
               | GaggiX wrote:
               | Are you sure you are not confusing nano banana pro for
               | nano banana, z-image still has a bit of AI look that I do
               | not find with nano banana pro, example for a comparison: 
               | https://i.ibb.co/YFtxs4hv/594068364-25101056889517041-340
               | 369...
               | 
               | Also Imagen 4 and Nano Banana Pro are very different
               | models.
        
               | cubefox wrote:
               | In your example, z-image and Nano Banana Pro look
               | basically equally photorealistic to me. Perhaps the NBP
               | image looks a bit more real because it resembles an
               | unstaged smartphone shot with wide angle. Anyway, the
               | difference is very small. I agree the lighting in Flux.2
               | Pro looks a bit off.
               | 
               | But anyway, realistic environments like a street cafe are
               | not suited to test for photorealism. You have to use
               | somewhat more fantastical environments.
               | 
               | I don't have access to z-image, but here are two examples
               | with Nano Banana Pro:
               | 
               | "A person in the streets of Atlantis, portrait shot."
               | https://i.ibb.co/DgMXzbxk/Gemini-Generated-
               | Image-7agf9b7agf9...
               | 
               | "A person in the streets of Atlantis, portrait shot
               | (photorealistic)" https://i.ibb.co/nN7cTzLk/Gemini-
               | Generated-Image-l1fm5al1fm5...
               | 
               | These are terribly unrealistic. Far more so than the
               | Flux.2 Pro image above.
               | 
               | > Also Imagen 4 and Nano Banana Pro are very different
               | models.
               | 
               | No, Imagen 4 is a pure diffusion model. Nano Banana Pro
               | is a Gemini scaffold which uses Imagen to generate an
               | initial image, then Gemini 3 Pro writes prompts to edit
               | the image for much better prompt alignment. The prompts
               | above a very simple, so there is little for Gemini to
               | alter, so they look basically identical to plain Imagen
               | 4. Both pictures (especially the first) have the
               | signature AI look of Imagen 4, which is different from
               | other models like Imagen 3.
               | 
               | By the way, here is GPT Image 1.5 with the same prompts:
               | 
               | "A person in the streets of Atlantis, portrait shot."
               | https://i.ibb.co/Df8nDHFL/Chat-GPT-
               | Image-10-Feb-2026-14-17-1...
               | 
               | "A person in the streets of Atlantis, portrait shot
               | (photorealistic)" https://i.ibb.co/Nns4pdGX/Chat-GPT-
               | Image-10-Feb-2026-14-17-2...
               | 
               | The first is very fake and the second is a strong
               | improvement, though still far from the excellent cafe
               | shots above (fake studio lighting, unrealistic colors
               | etc).
        
               | GaggiX wrote:
               | >In your example, z-image and Nano Banana Pro look
               | basically equally photorealistic to me
               | 
               | I disagree, nano banana pro result is on a completely
               | different league compare to flux.2 and z-image.
               | 
               | >But anyway, realistic environments like a street cafe
               | are not suited to test for photorealism
               | 
               | Why? It's the perfect settings in my opinion.
               | 
               | Btw I don't think you are using nano banana pro, probably
               | standard nano banana, I'm getting this from your prompt:
               | https://i.ibb.co/wZHx0jS9/unnamed-1.jpg
               | 
               | >Nano Banana Pro is a Gemini scaffold which uses Imagen
               | to generate an initial image, then Gemini 3 Pro writes
               | prompts to edit the image for much better prompt
               | alignment.
               | 
               | First of all how should you know the architecture details
               | of gemini-3-pro-image, second of all how the model can
               | modify the image if gemini itself is just rewriting the
               | prompt (like old chatgpt+dalle), imagen 4 is just a text-
               | to-image model, not an editing one, it doesn't make
               | sense, nano banana pro can edit images (like the ones you
               | can provide).
        
               | cubefox wrote:
               | > I disagree, nano banana pro result is on a completely
               | different league.
               | 
               | I strongly disagree. But even if you are right, the
               | difference between the cafe shots and the Atlantis shots
               | is clearly much, much larger than the difference between
               | the different cafe shots. The Atlantis shots are super
               | unrealistic. They look far worse than the cafe shots of
               | Flux.2 Pro.
               | 
               | > Why? It's the perfect settings in my opinion
               | 
               | Because it's too easy obviously. We don't need an AI to
               | make fake realistic photos of realistic environments when
               | we can easily photograph those ourselves. Unrealistic
               | environments are more discriminative because they are
               | much more likely to produce garbage that doesn't look
               | photorealistic.
               | 
               | > Btw I don't think you are using nano banana pro, I'm
               | getting this from your prompt:
               | https://i.ibb.co/wZHx0jS9/unnamed-1.jpg
               | 
               | I'm definitely using Nano Banana Pro, and your picture
               | has the same strong AI look to it that is typical of NBP
               | / Imagen 4.
               | 
               | > First of all how should you know the architecture
               | details of gemini-3-pro-image, second of all how the
               | model can modify the image if gemini itself is just
               | rewriting the prompt (like old chatgpt+dalle), imagen 4
               | is just a text-to-image model, not an editing one, it
               | doesn't make sense, nano banana pro can edit images (like
               | the ones you can provide).
               | 
               | There were discussions about it previously on HN. Clearly
               | NBP is using Gemini reasoning, and clearly the style of
               | NBP strongly resembles Imagen 4 specifically. There is
               | probably also a special editing model involved, just like
               | in Qwen-Imahe-2.0.
        
               | GaggiX wrote:
               | >Because it's too easy obviously.
               | 
               | Still the vast majority of models fail at delivery an
               | image that looks real, I want realism for a realistic
               | settings, if it can't do that than what's the point. Of
               | course you can always pay people and equipment to make
               | the perfect photo for you ahah
               | 
               | If the image of z-image turbo looks as good as the nano
               | banana pro one for you, you are probably too used to slop
               | that a model that do not produce obvious artifacts like
               | super shiny skin it's immediately undistinguishable from
               | a real image (like the nano banana pro one that to me
               | looks as real as a real photo) and yes I'm ignoring the
               | fact that in the z-image-turbo the cup is too large and
               | the bag is inside the chair. Z-image is good (in
               | particular given its size) but not as good.
        
               | cubefox wrote:
               | It seems you are ignoring the fact that the NBP Atlantis
               | pictures looks much, much worse than the z-image picture
               | of the cafe. They look far more like AI slop. (Perhaps
               | the Atlantis prompt would look even worse with z-image, I
               | don't know.)
        
               | GaggiX wrote:
               | I have generated my own using your prompt and post it in
               | the previous comment. You haven't posted a z-image one of
               | Atlantis. I'm not at home to try but I have trained lora
               | for z-image (it's a relatively lightweight model), I know
               | the model, it's not as good as nano banana pro. Use what
               | you prefer.
        
               | cubefox wrote:
               | > I have generated my own using your prompt and post it
               | in the previous comment.
               | 
               | Yes, and it has a very unrealistic AI look to it. That
               | was my point.
               | 
               | > You haven't posted a z-image one of Atlantis.
               | 
               | Yes, I don't doubt that it might well be just as
               | unrealistic or even worse. I also just tried the Atlantis
               | prompts in Grok (no idea what image model they use
               | internally) and they look somewhat more realistic, though
               | not on cafe level.
        
         | finnjohnsen2 wrote:
         | I agree. They makes me nauseous. The same kind of light nausea
         | as car sickness.
         | 
         | I assume our brains are used to stuff which we dont notice
         | conciously, and reject very mild errors. I've stared at the
         | picture a bit now and the finger holding the baloon is weird.
         | The out of place snowman feels weird. If you follow the
         | background blur around it isnt at the same depth everywehere.
         | Everything that reflects, has reflections that I cant see in
         | the scene.
         | 
         | I dont feel good staring at it now so I had to stop.
        
           | jbl0ndie wrote:
           | Sounds like you're describing the uncanny valley
           | https://en.wikipedia.org/wiki/Uncanny_valley
        
         | elorant wrote:
         | The lighting is wrong, that's what's telling to me. They look
         | too crisp. No proper shadows, everything looks crystal clear.
        
           | techpression wrote:
           | It's the HDR era all over again, where people edited their
           | photos to lack all contrast and just be ultra flat.
        
         | brookst wrote:
         | Everything is weightless. When real people stand and gesture
         | there's natural muscle use, hair and clothing drape, papers lay
         | flat on surfaces.
        
       | dsrtslnd23 wrote:
       | unfortunately no open weights it seems.
        
         | embedding-shape wrote:
         | To be fair, didn't they release open weights image model only
         | like a ~month ago? Think last one was in December 2025.
        
           | vunderba wrote:
           | Exactly - they did the same thing with the original version
           | of Qwen-Image. It was API only for a while before being made
           | available for local hosting.
        
       | raincole wrote:
       | It's crazy to think there was a fleeting sliver of time during
       | which Midjourney felt like the pinnacle of image generation.
        
         | Mashimo wrote:
         | What ever happend to midjourney?
        
           | raincole wrote:
           | Not much, while everything happened at OpenAI/Google/Chinese
           | companies. And that's the problem.
        
             | KeplerBoy wrote:
             | How is it a problem? There simply doesn't seem to be a moat
             | or secret sauce. Who cares which of these models is SOTA?
             | In two months there will be a new model.
        
               | waldarbeiter wrote:
               | There seems to be a moat like infrastructure/gpus and
               | talent. The best models right now come from companies
               | with considerable resources/funding.
        
               | esperent wrote:
               | Right, but that's a short term moat. If they pause on
               | their incredible levels of spending for even 6 months,
               | someone else will take over having spent only a tiny
               | fraction of what they did. They might get taken over
               | anyway.
        
               | raincole wrote:
               | > someone else will take over having spent only a tiny
               | fraction of what they did
               | 
               | How. By magic? You fell for 'Deepseek V3 is as good as
               | SOTA'?
        
               | Gud wrote:
               | By reverse engineering, sheer stupidity from the
               | competition, corporate espionage, 'stealing' engineers
               | and sometimes a stroke of genius, the same as it's always
               | been
        
           | wongarsu wrote:
           | They have image and video models that are nowhere near SOTA
           | on prompt adherence or image editing but pretty good on the
           | artistic side. They lean in on features like reference images
           | so objects or characters have a consistent look, biasing the
           | model towards your style preferences, or using moodboards to
           | generate a consistent style
        
           | Lalabadie wrote:
           | No external funding raised. They're not on the VC path, so no
           | need to chase insane growth. They still have around 500M USD
           | in ARR.
           | 
           | In my (very personal) opinion, they're part of a very small
           | group of organizations that sell inference under a sane and
           | successful business model.
        
             | spaceman_2020 wrote:
             | Aesthetically, still unmatched
        
             | aenvoker wrote:
             | Not on the VC path. Not even on the max-profit path. Just
             | on the "Have fun doing cool research" path.
             | 
             | I was a mod on MJ for its first few years and got to know
             | MJ's founder through discussions there. He already had
             | "enough" money for himself from his prior sale of Leap
             | Motion to do whatever he wanted. And, he decided what he
             | wanted was to do cool research with fun people. So, he
             | started MJ. Now he has far more money than before and what
             | he wants to do with it is to have more fun doing more cool
             | research.
        
             | echelon wrote:
             | They're working on a few really lofty ideas:
             | 
             | 1. real time world models for the "holodeck". It has to be
             | fast, high quality, and inexpensive for lots of users. They
             | started on this two years ago before "world model" hype was
             | even a thing.
             | 
             | 2. some kind of hardware to support this.
             | 
             | David Holz talks about this on Twitter occasionally.
             | 
             | Midjourney still has incredible revenue. It's still the
             | best looking image model, even if it's hard to prompt,
             | can't edit, and has artifacting. Every generation looks
             | like it came out of a magazine, which is something the
             | other leading commercial models lack.
        
           | vunderba wrote:
           | A lot of people started realizing that it didn't really
           | matter how pretty the resulting image was if it completely
           | failed to adhere to the prompt.
           | 
           | Even something like Flux.1 Dev which can be run entirely
           | locally and was released back in August of 2024 has
           | significantly better prompt understanding.
        
           | qingcharles wrote:
           | They still have a niche. Their style references feature is
           | their key differentiator now, but I find I can usually just
           | drop some images of a MJ style into Gemini and get it to give
           | me a text prompt that works just as well as MJ srefs.
        
         | gamma-interface wrote:
         | The pace of commoditization in image generation is wild. Every
         | 3-4 months the SOTA shifts, and last quarter's breakthrough
         | becomes a commodity API.
         | 
         | What's interesting is that the bottleneck is no longer the
         | model -- it's the person directing it. Knowing what to ask for
         | and recognizing when the output is good enough matters more
         | than which model you use. Same pattern we're seeing in code
         | generation.
        
           | SV_BubbleTime wrote:
           | SOTA shifts, yes. But the average person doing the work has
           | been very happy with SDXL based models. And that was released
           | two years ago.
           | 
           | The fight right now outside of API SOTA is who will replace
           | SDXL to be the "community preference"
           | 
           | It's now a three way between Flux2 Klein, Z-Image, and now
           | Qwen2.
        
           | echelon wrote:
           | I'm happy the models are becoming commodity, but we still
           | have a long way to go.
           | 
           | I want the ability to lean into any image and tweak it like
           | clay.
           | 
           | I've been building open source software to orchestrate the
           | frontier editing models (skip to halfway down), but it would
           | be nice if the models were built around the software
           | manipulation workflows:
           | 
           | https://getartcraft.com/news/world-models-for-film
        
           | sincerely wrote:
           | PLEASE STOP POSTING AI GENERATED COMMENTS
        
         | rc1 wrote:
         | Isn't it still? Antidotally, I work with lots of creators who
         | still prefer it because of its subjective qualities.
        
       | skerit wrote:
       | > Qwen-Image-2.0 not only accurately models the "riding" action
       | but also meticulously renders the horse's musculature and hair >
       | https://qianwen-res.oss-accelerate-overseas.aliyuncs.com/Qwe...
       | 
       | What the actual fuck
        
         | wongarsu wrote:
         | For reference, below is the prompt translated (with my
         | highlighting of the part that matters). They did very much ask
         | for this version of "horse riding a man", not the "horse
         | sitting upright on a crawling human" version
         | 
         | ---
         | 
         | A desolate grassland stretches into the distance, its ground
         | dry and cracked. Fine dust is kicked up by vigorous activity,
         | forming a faint grayish-brown mist in the low sky.
         | 
         | Mid-ground, eye-level composition: _A muscular, robust adult
         | brown horse stands proudly, its forelegs heavily pressing
         | between the shoulder blades and spine of a reclining man_. Its
         | hind legs are taut, its neck held high, its mane flying against
         | the wind, its nostrils flared, and its eyes sharp and focused,
         | exuding a primal sense of power. _The subdued man is a white
         | male, 30-40 years old_ , his face covered in dust and sweat,
         | his short, messy dark brown hair plastered to his forehead, his
         | thick beard slightly damp; he wears a badly worn, grey-green
         | medieval-style robe, the fabric torn and stained with mud in
         | several places, a thick hemp rope tied around his waist, and
         | scratched ankle-high leather boots; _his body is in a push-up
         | position_ --his palms are pressed hard against the cracked, dry
         | earth, his knuckles white, the veins in his arms bulging, his
         | legs stretched straight back and taut, his toes digging into
         | the ground, his entire torso trembling slightly from the
         | weight.
         | 
         | The background is a range of undulating grey-blue mountains,
         | their outlines stark, their peaks hidden beneath a low-hanging,
         | leaden-grey, cloudy sky. The thick clouds diffuse a soft,
         | diffused light, which pours down naturally from the left front
         | at a 45-degree angle, casting clear and voluminous shadows on
         | the horse's belly, the back of the man's hands, and the cracked
         | ground.
         | 
         | The overall color scheme is strictly controlled within the
         | earth tones: the horsehair is warm brown, the robe is a
         | gradient of gray-green-brown, the soil is a mixture of ochre,
         | dry yellow earth, and charcoal gray, the dust is light
         | brownish-gray, and the sky is a transition from matte lead gray
         | to cool gray with a faint glow at the bottom of the clouds.
         | 
         | The image has a realistic, high-definition photographic
         | quality, with extremely fine textures--you can see the sweat on
         | the horse's neck, the wear and tear on the robe's warp and weft
         | threads, the skin pores and stubble, the edges of the cracked
         | soil, and the dust particles. The atmosphere is tense,
         | primitive, and full of suffocating tension from a struggle of
         | biological forces.
        
           | badhorseman wrote:
           | The significance of the hemp-rope is that it is symbol of
           | morning and loss of ones decedent.
        
         | embedding-shape wrote:
         | I like how sometimes I get angry at a LLM for not understanding
         | what I meant, but then I realize that I just forgot to mention
         | it in the context. It's fun to see the same thing happen in
         | humans reading websites too, where they don't understand the
         | context yet react with strong feelings anyways.
        
       | cocodill wrote:
       | interesting riding application picture
        
         | rwmj wrote:
         | "Guy being humped by a horse" wouldn't have been my first
         | choice for demoing the capabilities of the model, but each to
         | their own I guess.
        
           | viraptor wrote:
           | It looks like a marketing move. It's a good quality, detailed
           | picture. It's going to get shared a lot. I would assume they
           | knew exactly what they were doing. Nothing like a bit of
           | controversy for extra clicks.
        
             | brookst wrote:
             | Because every ML researcher is a viral social media expert.
             | 
             | (I don't even know if I'm being sarcastic)
        
               | viraptor wrote:
               | This is not some random ML researcher doing fun things at
               | home. Qwen is backed by Alibaba cloud. They likely have
               | whole departments of marketing people available.
        
       | fguerraz wrote:
       | I found the horse revenge-porn image at the end quite disturbing.
        
         | embedding-shape wrote:
         | I think they call it "horse riding a human" which could have
         | taken two very different directions, and the direction the
         | model seems to have taken was the least worst of the two.
        
           | wongarsu wrote:
           | At first I thought it's a clever prompt because you see which
           | direction the model takes it, and whether it "corrects" it to
           | the more common "human riding a horse" similar to the full
           | wine glass test.
           | 
           | But if you translate the actual prompt the term riding
           | doesn't even appear. The prompt describes the exact thing you
           | see in excruciating detail.
           | 
           | "... A muscular, robust adult brown horse standing proudly,
           | its forelegs heavily pressing between the shoulder blades and
           | spine of a reclining man ... and its eyes sharp and focused,
           | exuding a primal sense of power. The subdued man is a white
           | male, 30-40 years old, his face covered in dust and sweat ...
           | his body is in a push-up position--his palms are pressed hard
           | against the cracked, dry earth, his knuckles white, the veins
           | in his arms bulging, his legs stretched straight back and
           | taut, his toes digging into the ground, his entire torso
           | trembling slightly from the weight ..."
        
             | embedding-shape wrote:
             | > But if you translate the actual prompt the term riding
             | doesn't even appear. The prompt describes the exact thing
             | you see in excruciating detail.
             | 
             | Yeah, as they go through their workflow earlier in the blog
             | post, that prompt they share there seems to be generated by
             | a different input, then that prompt is passed to the actual
             | model. So the workflow is something like "User prompt input
             | -> Expand input with LLMs -> Send expanded prompt to image
             | model".
             | 
             | So I think "human riding a horse" is the user prompt, which
             | gets expanded to what they share in the post, which is what
             | the model actually uses. This is also how they've presented
             | all their previous image models, by passing user input
             | through a LLM for "expansion" first.
             | 
             | Seems poorly thought out not to make it 100% clear what the
             | actual humanly-written prompt is though, not sure why they
             | wouldn't share that upfront.
        
           | chakintosh wrote:
           | Is it related to "Mr Hands" ?
        
         | blitzar wrote:
         | Wont someone think of the horses.
        
         | engcoach wrote:
         | It's the year of the horse in their zodiac. The (translated)
         | prompt is wild:
         | 
         | """ A desolate grassland stretches into the distance, its
         | ground dry and cracked. Fine dust is kicked up by vigorous
         | activity, forming a faint grayish-brown mist in the low sky.
         | Mid-ground, eye-level composition: A muscular, robust adult
         | brown horse stands proudly, its forelegs heavily pressing
         | between the shoulder blades and spine of a reclining man. Its
         | hind legs are taut, its neck held high, its mane flying against
         | the wind, its nostrils flared, and its eyes sharp and focused,
         | exuding a primal sense of power. The subdued man is a white
         | male, 30-40 years old, his face covered in dust and sweat, his
         | short, messy dark brown hair plastered to his forehead, his
         | thick beard slightly damp; he wears a badly worn, grey-green
         | medieval-style robe, the fabric torn and stained with mud in
         | several places, a thick hemp rope tied around his waist, and
         | scratched ankle-high leather boots; his body is in a push-up
         | position--his palms are pressed hard against the cracked, dry
         | earth, his knuckles white, the veins in his arms bulging, his
         | legs stretched straight back and taut, his toes digging into
         | the ground, his entire torso trembling slightly from the
         | weight. The background is a range of undulating grey-blue
         | mountains, their outlines stark, their peaks hidden beneath a
         | low-hanging, leaden-grey, cloudy sky. The thick clouds diffuse
         | a soft, diffused light, which pours down naturally from the
         | left front at a 45-degree angle, casting clear and voluminous
         | shadows on the horse's belly, the back of the man's hands, and
         | the cracked ground. The overall color scheme is strictly
         | controlled within the earth tones: the horsehair is warm brown,
         | the robe is a gradient of gray-green-brown, the soil is a
         | mixture of ochre, dry yellow earth, and charcoal gray, the dust
         | is light brownish-gray, and the sky is a transition from matte
         | lead gray to cool gray with a faint glow at the bottom of the
         | clouds. The image has a realistic, high-definition photographic
         | quality, with extremely fine textures--you can see the sweat on
         | the horse's neck, the wear and tear on the robe's warp and weft
         | threads, the skin pores and stubble, the edges of the cracked
         | soil, and the dust particles. The atmosphere is tense,
         | primitive, and full of suffocating tension from a struggle of
         | biological forces. """
        
       | cubefox wrote:
       | The complex prompt following ability and editing is seriously
       | impressive here. They don't seem to be much behind OpenAI and
       | Google. Which is backed op by the AI Arena ranking.
        
       | yieldcrv wrote:
       | when the horsey tranq hits
        
       | wiether wrote:
       | I use gen-AI to produce images daily, but honestly the
       | infographics are 99% terrible.
       | 
       | LinkedIn is filled with them now.
        
         | smcleod wrote:
         | To be fair it hasn't made LinkedIn any worse than it already
         | was.
        
           | nurettin wrote:
           | To be fair, it is hard to make LinkedIn any worse.
        
             | embedding-shape wrote:
             | I was gonna make a joke about "Wish granted, now Microsoft
             | owns it" but then I remembered that they already do.
             | Reality sometimes makes better jokes than what we can come
             | up with.
        
             | aenis wrote:
             | And at the same time, its arguably the least toxic of all
             | social networks.
             | 
             | Yes, cringeworthy but at least not addictive! Its like
             | facebook all those years ago, i can IM friends from
             | highschool without having to pay any attention to the feed.
        
         | usefulposter wrote:
         | Correct.
         | 
         | Much like the pointless ASCII diagrams in GitHub readmes (big
         | rectangle with bullet points flows to another...), the diagrams
         | are cognitive slurry.
         | 
         | See Gas Town for non-Qwen examples of how bad it can get:
         | 
         | https://news.ycombinator.com/item?id=46746045
         | 
         | (Not commenting on the other results of this model outside of
         | diagramming.)
        
           | viraptor wrote:
           | > cognitive slurry
           | 
           | Thank you for this phrase. I don't think that bad diagrams
           | are limited to the AI in any way and this perfectly describes
           | all "this didn't make things any clearer" cases.
        
         | viraptor wrote:
         | Informatics are as bad as the author allows though. There's few
         | people who could make or even describe a good infographic, so
         | that's what we see in the results too.
        
         | mdrzn wrote:
         | Infographics and full presentations are a NanoBananaPro
         | exclusive so far.
        
           | RationPhantoms wrote:
           | You should see some of the work from their PaperBanana
           | papers. Really solid.
        
       | inanothertime wrote:
       | I recently tried out LMStudio on Linux for local models. So easy
       | to use!
       | 
       | What Linux tools are you guys using for image generation models
       | like Qwen's diffusion models, since LMStudio only supports text
       | gen.
        
         | ilaksh wrote:
         | I have my own MIT licensed framework/UI:
         | https://github.com/runvnc/mindroot. With Nano Banana via
         | runvnc/googleimageedit
        
         | guai888 wrote:
         | ComfyUI is the best for stable diffusion
        
           | embedding-shape wrote:
           | FWIW you can use non-sd models in ComfyUI too, the ecosystem
           | is pretty huge and supports most of the "mainstream" models,
           | not only the stable diffusion ones, even video models and
           | more too.
        
         | embedding-shape wrote:
         | Everything keeps changing so quickly, I basically have my own
         | Python HTTP server with a unified JSON interface, then that can
         | be routed to any of the impls/*.py files for the actual
         | generation, then I have of those per
         | implementation/architecture basically. Mostly using `diffusers`
         | for the inference, which isn't the fastest, but tends to have
         | the new model architectures much sooner than everyone else.
        
         | PaulKeeble wrote:
         | Ollama is working on adding image generation but its not here
         | yet. We really do need something that can run a variety of
         | models for images.
        
           | embedding-shape wrote:
           | Yeah, I'm guessing they were bound to leave behind the whole
           | "Get up and running with large language models" mission
           | sooner or later, which was their initial focus, as investors
           | after 2-3 years start making you to start thinking about
           | expansion and earning back the money.
           | 
           | Sad state of affairs and seems they're enshittifying quicker
           | than expected, but was always a question of when, not if.
        
         | eurekin wrote:
         | Practically anybody actually creating with this class of models
         | (diffusion based mostly) is using ComfyUI. Community takes care
         | of quantization, repackaging into gguf (most popular) and even
         | speed optimizing (lighting loras, layers skip). It's quite
         | extensive
        
         | sequence7 wrote:
         | If you're on an AMD platform Lemonade (https://lemonade-
         | server.ai/) added image generation in version 9.2
         | (https://github.com/lemonade-sdk/lemonade/releases/tag/v9.2.0).
        
         | vunderba wrote:
         | I encourage everyone to at least try ComfyUI. It's come a long
         | way in terms of user-friendliness particularly with all of the
         | built-in Templates you can use.
        
         | Eisenstein wrote:
         | Koboldcpp has built in support for image models. Model search
         | and download, one executable to run, UI, OpenAI API endpoint,
         | llama.cpp endpoint, highly configurable. If you want to get up
         | and running instantly, just pick a kcppt file and open that and
         | it will download everything you need and load it for you.
         | 
         | Engine:
         | 
         | * https://github.com/LostRuins/koboldcpp/releases/latest/
         | 
         | Kcppt files:
         | 
         | * https://huggingface.co/koboldcpp/kcppt/tree/main
        
         | SV_BubbleTime wrote:
         | LMStudio is a low barrier to entry for LLMs, for sure. The
         | lowest. Good software!
         | 
         | Other people gave you the right answer, ComfyUI. I'll give you
         | the more important why and how...
         | 
         | There is a huge effort of people to do everything but Comfy
         | because of its intimidating barrier. It's not that bad. Learn
         | it once and be done. You won't have to keep learning UI of the
         | week endlessly.
         | 
         | The how, go to civitai. Find an image you like, drag and drop
         | it into comfy. If it has a workflow attached, it will show you.
         | Install any missing nodes they used. Click the loaders to point
         | to your models instead of their models. Hit run and get the
         | same or a similar image. You don't need to know what any of the
         | things do yet.
         | 
         | If for some reason that just does not work for you... Swarm UI,
         | is a front end too comfy. You can change things and it will
         | show you on the comfy side what they're doing. It's a gateway
         | drug to learning comfy.
         | 
         | EDIT: most important thing no one will tell you out right... DO
         | NOT FOR ANY REASON try and skip the VENV or miniconda virtual
         | environment when using comfy! You must make a new and clean
         | setup. You will never get the right python, torch, diffusers,
         | driver, on your system install.
        
         | adammarples wrote:
         | Stability matrix, it's a manager for models and uis and loras
         | etc, very nice
        
       | goga-piven wrote:
       | Why is the only image featuring non-Asian men the one under the
       | horse?
        
         | andruby wrote:
         | Is the problem the position/horse or that Qwen mostly shows
         | asian people?
         | 
         | Do western AI models mostly default to white people?
        
           | goga-piven wrote:
           | Well, what if some western models showcase white people in
           | all good-looking images and the only embarrassing image
           | features Asian people? wouldn't that be considered racism?
        
             | embedding-shape wrote:
             | > and the only embarrassing image
             | 
             | Embarrassing image? I'm white, why would I be embarrassed
             | over that image? It's a computer generated image with no
             | real people in it, how could it be embarrassing for alive
             | humans?
        
         | z3dd wrote:
         | they explicitly called for that in the prompt
        
           | wtcactus wrote:
           | But why? That image actually puzzled me. Does it have some
           | background context? Some historical legend or something of
           | the like?
        
             | joeycodes wrote:
             | It is Lunar New Year season right now, 2026 is year of the
             | horse, there is celebratory horse imagery everywhere in
             | many Asian countries right now, so this image could be
             | interpreted as East trampling West. I have no way to know
             | the intention of the person at Qwen who wrote this, but you
             | can form your own conclusions from the prompt:
             | 
             |  _A muscular, robust adult brown horse stands proudly, its
             | forelegs heavily pressing between the shoulder blades and
             | spine of a reclining man. Its hind legs are taut, its neck
             | held high, its mane flying against the wind, its nostrils
             | flared, and its eyes sharp and focused, exuding a primal
             | sense of power. The subdued man is a white male..._
        
           | goga-piven wrote:
           | Exactly why did they choose this prompt with a white person
           | and not an Asian person, as in all the other examples?
        
       | singularfutur wrote:
       | Another closed model dressed up as "coming soon" open source. The
       | pattern is obvious: generate hype with a polished demo, lock the
       | weights, then quietly move on. Real open source doesn't need a
       | press release countdown.
        
         | kkzz99 wrote:
         | Good that we have the arbitrator of what "real open source" is
         | and isn't over here.
        
           | yellowapple wrote:
           | "Open source" is indeed an objective standard with actual
           | criteria, and not just vibes.
           | 
           | Luckily, it seems previous Qwen models did get open-sourced
           | in the actual sense, so this one probably will be, too.
        
             | kkzz99 wrote:
             | Its one NGO trying to dictate what "open source" is and
             | btw. according to that definition, Qwen isn't open source.
        
         | yorwba wrote:
         | Where do you see a press release countdown? Alibaba
         | consistently doesn't release weights for their biggest models,
         | but they also don't pretend that they do.
        
           | singularfutur wrote:
           | fair enough, I read too quickly the article.
        
         | vunderba wrote:
         | That's not what they did with Qwen-Image v1 - they announced it
         | and it was available via API, but then they released the
         | weights a few weeks after with an Apache 2.0 license. Let's at
         | least give them the benefit of the doubt here.
        
       | sandbach wrote:
       | The Chinese vertical typography is sadly a bit off. If
       | punctuation marks are used at all, they should be the characters
       | specifically designed for vertical text, like [?](U+FE12
       | PRESENTATION FORM FOR VERTICAL IDEOGRAPHIC FULL STOP).
        
       | tianqi wrote:
       | I've seen many comments describing the "horse riding man" example
       | as extremely bizarre (which it actually is), so I'd like to
       | provide some background context here. The "horse riding man" is a
       | Chinese internet meme originating from an entertainment awards
       | ceremony, when the renowned host Tsai Kang-yong wore an elaborate
       | outfit featuring a horse riding on his back[1]. At the time, he
       | was embroiled in a rumor about his unpublicized homosexual
       | partner, whose name sounded "Ma Qi Ren" which coincidentally
       | translates to "horse riding man" in Mandarin. This incident
       | spread widely across Chinese internet and turned into a meme. So
       | they used "horse riding man" as an example isn't entirely
       | nonsensical, though the image per se is undeniably bizarre and
       | carries an unsettling vibe.
       | 
       | [1] The photo of the outfit:
       | https://share.google/mHJbchlsTNJ771yBa
        
         | badhorseman wrote:
         | Why not ask for simply a man or even an Han man given the race
         | of Tsai Kang-yong. Why a white man and why a man wearing
         | medieval clothing. Gives your head a wobble.
        
           | DustinEchoes wrote:
           | Yep, it's the only image on the entire page with a non-
           | Chinese person in it. Given the prompt, the message is clear.
        
             | yorwba wrote:
             | The message is "We watched Lord of the Rings and Game of
             | Thrones and liked the medieval aesthetic enough to emulate
             | it."
        
               | badhorseman wrote:
               | remind me of the bit of lord of the rings where muscular
               | horses dominate European peasant men, as per the prompt
               | translation.
        
               | yorwba wrote:
               | Yes, in those movies, the hot white guys (and sometimes
               | girls) usually ride on top of the muscular horses. So
               | when you want to show a horse riding a man as a visual
               | gag, why not make the man a hot white guy with a gruff
               | beard?
               | 
               | You act as though they first decided to make an image
               | representing Westerners and then chose that particular
               | scene as an intentional insult, but you need to consider
               | that they likely made thousands of test images, most of
               | which were just playing around with the model's
               | capabilities and not specifically crafted for the
               | announcement post.
               | 
               | So why did this one get picked? I think it boils down to
               | the visual gag being funny and the movie-like quality.
        
               | popalchemist wrote:
               | Racial/cultural tension is part of the context in which
               | this image is appearing. Not only because of historical
               | tensions, but because this image appears as part of this
               | generation's Manhattan Project style arms race toward AGI
               | and global dominance. Your denial of that is a reflection
               | of your own ignorance.
        
         | yorwba wrote:
         | There's also the "horse riding astronaut" challenge in image
         | generation: https://garymarcus.substack.com/p/horse-rides-
         | astronaut-redu...
        
           | laughingcurve wrote:
           | Gary Marcus is not the man to be looking to on this topic
        
             | AlphaAndOmega0 wrote:
             | Gary Marcus successfully predicted all ten of the one AI
             | Winters.
             | 
             | He also claimed that LLMs were a failure because of prompts
             | that GPT 3.5 couldn't parse, _after_ the launch of
             | GPT-4,which handled them with aplomb.
        
             | yorwba wrote:
             | Gary Marcus successfully wrote an article about getting
             | image generation models to show a horse riding an
             | astronaut, which is all I needed him to do. (Actually he
             | wrote two, but this one felt more concise.) Take it as an
             | existence proof, not an endorsement.
        
               | observationist wrote:
               | Just like a basilisk, if you never refer to him again, he
               | fades away and doesn't bug people anymore. Let him fight
               | through whatever he needs to if he ever bothers coming up
               | with anything the rest of the world needs to hear; until
               | then, we can enjoy the peace and quiet.
        
         | Lerc wrote:
         | On the topic of modern Chinese culture, Is there the same
         | hostility towards AI generated Imagery in China as there seems
         | to be in America?
         | 
         | For example I think there would be a lot of businesses in the
         | US that would be too afraid of backlash to use AI generated
         | imagery for an itinerary like the one at https://qianwen-
         | res.oss-accelerate-overseas.aliyuncs.com/Qwe...
        
           | tianqi wrote:
           | Since China has a population of 1.4 billion people with
           | vastly differing levels of cognition, I find it difficult to
           | claim I can summarize "modern Chinese culture". But within my
           | range of observation, no. Chinese not only have no hostility
           | toward AI but actively pursues and reveres it with fervor.
           | They widely perceive AI as an advanced force, a new
           | opportunity for everyone, a new avenue for making money, and
           | a new chance to surpass others. At most, some of the
           | consumers might associate businesses using AI generated
           | content with a budget-conscious brand image, but not hostile.
        
             | Lerc wrote:
             | _> Since China has a population of 1.4 billion people with
             | vastly differing levels of cognition, I find it difficult
             | to claim I can summarize "modern Chinese culture"_
             | 
             | Ha! An American would have no such qualms.
        
               | idiotsecant wrote:
               | Well, they would modify it slightly to claim " _real_
               | American culture is... " In general, the range of 'real'
               | America is about 300 miles, in my experience.
        
           | yorwba wrote:
           | There's definitely _some_ hostility:
           | https://mp.weixin.qq.com/s/A5shO-6nZIXZvJUEzrx03Q
        
             | yieldcrv wrote:
             | > "Why did such a strange metaphor like _' the sound of an
             | electrocardiogram machine moving paper'_ appear in this
             | story that had nothing to do with medicine?"
             | 
             | this is sending me, I don't know what's funnier, this
             | translation being accurate or inaccurate
        
         | vessenes wrote:
         | Interesting background! Prompts like this also test the latent
         | space of the image generator - it's usually the other way
         | round, so if you see a man on top of a horse, you've got a less
         | sophisticated embedding feeding the model. In this case,
         | though, that's quite an image to put out to the interwebs. I
         | looked to see what gender the horse was.
         | 
         | EDIT: After reading the prompt translation, this was more just
         | like a "year of the horse is going to nail white engineers in
         | glorious rendered detail" sort of prompt. I don't know how
         | SD1.5 would have rendered it, and I think I'll skip finding out
        
         | cubefox wrote:
         | While I don't doubt this was one influence, there was also an
         | infamous problem with Dall-E 2, which was perfectly able to
         | generate an astronaut riding a horse but completely unable to
         | generate a horse riding an astronaut.
         | 
         | This problem is infamous because it persisted (unlike other
         | early problems, like creating the wrong number of fingers) for
         | much more capable models, and the Qwen Image people are
         | certainly very aware of this difficult test. Even Imagen 4
         | Ultra, which might be the most advanced pure diffusion model
         | without editing loop, fails at it.
         | 
         | And obviously an astronaut is similar to a man, which connects
         | this benchmark to the Chinese meme.
        
         | bogzz wrote:
         | Fun fact, the Serbian parliament building has two statues of
         | horses riding men in front of it.
         | 
         | Which is really apt because in Serbian "konj", or horse, is a
         | colloquial word for moron. So, horses riding people is a
         | perfect representation of the reality of the Serbian
         | government.
         | 
         | Another fun fact, the parliament building in HL2's City 17 was
         | modelled from that building.
        
         | rahimnathwani wrote:
         | This is fascinating!
         | 
         | From the article it seems the name is Ma Qi Ren , not Ma Qi Ren
         | so the guy's name _sounds_ the same as  'horse riding man', but
         | that's not a literal translation of his name.
        
         | TacticalCoder wrote:
         | Very interesting! What's weird though is that the chinese do
         | not even pretend: every single picture has asian-looking people
         | generated.
         | 
         | But on the _one_ picture that honestly looks like a man getting
         | ass-raped by a horse, it 's a white man.
         | 
         | I mean even in the west where you can hardly see an ad with a
         | white couple anymore, they don't go that far (at least not
         | yet).
         | 
         | White people are a minority on earth and anti-white racism sure
         | seems to be alive and well (btw my family is of all the colors
         | and we speak three languages at home, so don't even try me).
        
         | popalchemist wrote:
         | Super tone-deaf and inappropriate. Not realizing how it would
         | read to the uninformed is a bad look. Myopic.
        
       | engcoach wrote:
       | The "horse riding man" prompt is wild:
       | 
       | """A desolate grassland stretches into the distance, its ground
       | dry and cracked. Fine dust is kicked up by vigorous activity,
       | forming a faint grayish-brown mist in the low sky. Mid-ground,
       | eye-level composition: A muscular, robust adult brown horse
       | stands proudly, its forelegs heavily pressing between the
       | shoulder blades and spine of a reclining man. Its hind legs are
       | taut, its neck held high, its mane flying against the wind, its
       | nostrils flared, and its eyes sharp and focused, exuding a primal
       | sense of power. The subdued man is a white male, 30-40 years old,
       | his face covered in dust and sweat, his short, messy dark brown
       | hair plastered to his forehead, his thick beard slightly damp; he
       | wears a badly worn, grey-green medieval-style robe, the fabric
       | torn and stained with mud in several places, a thick hemp rope
       | tied around his waist, and scratched ankle-high leather boots;
       | his body is in a push-up position--his palms are pressed hard
       | against the cracked, dry earth, his knuckles white, the veins in
       | his arms bulging, his legs stretched straight back and taut, his
       | toes digging into the ground, his entire torso trembling slightly
       | from the weight. The background is a range of undulating grey-
       | blue mountains, their outlines stark, their peaks hidden beneath
       | a low-hanging, leaden-grey, cloudy sky. The thick clouds diffuse
       | a soft, diffused light, which pours down naturally from the left
       | front at a 45-degree angle, casting clear and voluminous shadows
       | on the horse's belly, the back of the man's hands, and the
       | cracked ground. The overall color scheme is strictly controlled
       | within the earth tones: the horsehair is warm brown, the robe is
       | a gradient of gray-green-brown, the soil is a mixture of ochre,
       | dry yellow earth, and charcoal gray, the dust is light brownish-
       | gray, and the sky is a transition from matte lead gray to cool
       | gray with a faint glow at the bottom of the clouds. The image has
       | a realistic, high-definition photographic quality, with extremely
       | fine textures--you can see the sweat on the horse's neck, the
       | wear and tear on the robe's warp and weft threads, the skin pores
       | and stubble, the edges of the cracked soil, and the dust
       | particles. The atmosphere is tense, primitive, and full of
       | suffocating tension from a struggle of biological forces."""
        
         | badhorseman wrote:
         | some context for the perplexed.
         | https://live2makan.com/2024/08/07/treasures-statue-of-horse-...
        
       | vunderba wrote:
       | Couple of thoughts:
       | 
       | 1. I'd wager that given their previous release history, this will
       | be open-weight within 3-4 weeks.
       | 
       | 2. It looks like they're following suit with other models like
       | Z-Image Turbo (6B parameters) and Flux.2 Klein (9B parameters),
       | aiming to release models that can run on much more modest GPUs.
       | For reference, the original Qwen-Image is a 20B-parameter model.
       | 
       | 3. This is a unified model (both image generation and editing),
       | so there's no need to keep separate Qwen-Image and Qwen-Edit
       | models around.
       | 
       | 4. The original Qwen-Image scored the highest among local models
       | for image editing in my GenAI Showdown (6 out of 12 points), and
       | it also ranked very highly for image generation (4 out of 12
       | points).
       | 
       |  _Generative Comparisons of Local Models:_
       | 
       | https://genai-showdown.specr.net/?models=fd,hd,kd,qi,f2d,zt
       | 
       |  _Editing Comparison of Local Models:_
       | 
       | https://genai-showdown.specr.net/image-editing?models=kxd,og...
       | 
       | I'll probably be waiting until the local version drops before
       | adding Qwen-Image-2 to the site.
        
         | SV_BubbleTime wrote:
         | For the more technical...
         | 
         | Qwen 2512 (December edition of Qwen Image)
         | 
         | * 19B parameters, which was a 40GB file at FP16 and fit on a
         | 3090 at FP8. Anything less than that and you were in GGUF
         | format at Q6 to Q4 quantizations... which were slow, but still
         | good quality.
         | 
         | * used Qwen 2.5 VL. So a large model and a very good vision
         | model.
         | 
         | * And iirc, their own VAE. Which had known and obvious issues
         | of high frequency artifacts. Some people would take the image
         | and pass it through another VAE like WAN Video model's or
         | upscale-downscale to remove these
         | 
         | Qwen 2 now is
         | 
         | * a 7B param model. Right between Klein 9B (non-commercial)
         | this (license unknown), Z-Image 7B (Apache), and Klein 4B
         | (Apache). Direct competition, will fit on many more GPUs even
         | at FP16.
         | 
         | * upgrades to Qwen 3 VL, I assume this is better than the
         | already great 2.5 VL.
         | 
         | * Unknown on the new VAE. Flux2's new 128 channel VAE is
         | excellent, but it hasn't been out long enough for even a
         | frontier Chinese model to pick up.
         | 
         | Overall, you're right this is on the trend to bring models on
         | to lower end hardware.
         | 
         | Qwen was already excellent and now they rolled Image and Edit
         | together for an "Omni" model.
         | 
         | Z-Image was the model to beat a couple weeks ago... and now it
         | looks like both Klein and Qwen will! Z-Image has been
         | disappointing to see how it just refuses to adhere to multiple
         | new training concepts. Maybe they tried to pack it too tightly.
         | 
         | Open weights for this will be amazing. THREE direct competitors
         | all vying to be "SDXL2" at the same time.
         | 
         | The Qwen convention was confusing! You had Image, 2509, Edit,
         | 2511 (Edit), 2512 (Image) and then the Lora compatibility was
         | unspecified. It's smart to just 2.0 this mess.
        
           | liuliu wrote:
           | Note that Qwen Image 1.0 (2512) wasted ~8B weights on
           | timestep embedding. Both Z-Image / FLUX.2 series corrected
           | that.
        
           | vunderba wrote:
           | Agreed! A lot of people were also using ZiT as a refiner
           | downstream to help with some of the more problematic visual
           | aspects of the original Qwen-Image.
           | 
           | I'm really looking forward to running the unified model
           | through its paces.
        
             | SV_BubbleTime wrote:
             | Something I am skeptical about Z-Image is that it uses
             | Gemma which is imo a weak LLM.
             | 
             | If I were to guess, I would say that Z-Image's life is
             | shorter than it initially appeared. Even as a refiner which
             | are just workarounds for model issues.
        
       | wtcactus wrote:
       | So, I've just gave it this prompt:
       | 
       |  _" Analyze this webpage:
       | https://en.wikipedia.org/wiki/1989_Tiananmen_Square_protests...
       | 
       | Generate an infographic with all the data about the main event
       | timeline and estimated number of victims.
       | 
       | The background image should be this one:
       | https://en.wikipedia.org/wiki/Tank_Man#/media/File
       | :Tank_Man_(Tiananmen_Square_protester).jpg
       | 
       | Improve the background image clarity and resolution."_
       | 
       | I've received an error:
       | 
       |  _" Oops! There was an issue connecting to Qwen3-Max. Content
       | Security Warning: The input file data may contain inappropriate
       | content."_
       | 
       | I wonder if locally running the model they published in December
       | does have the same censorship in place (i.e. if it's already
       | trained like this), or if they implement the censorship by the
       | Chinese regimen in place for the web service only.
        
       | engcoach wrote:
       | My response to the horse image:
       | https://i.postimg.cc/hG8nJ4cv/IMG-5289-copy.jpg
        
       | ranger_danger wrote:
       | When I tried Qwen-Image-2512 I could not even get it to spell
       | correctly. And often the letters would be garbled anyways.
        
       | thisisit wrote:
       | I liked their comic panels example and tried it using their chat
       | at: https://chat.qwen.ai/
       | 
       | When I used the exact prompt the post - the chat works. It gives
       | me the exact output from the blog post.
       | 
       | Then I used Google Translate to understand the prompt format. The
       | prompt is: _A 4x6 panel comic, four lines, six panels per line.
       | Each panel is separated by a white dividing line.
       | 
       | The first row, from left to right: Panel 1: Panel 2: ....._
       | 
       | and when I try to change the inputs the comic example fails
       | miserably. It keeps creating random grids - sometimes 4x5 other
       | times 4x6 but then by third row the model will get confused and
       | the output has only 3 panels. Other times English dialogue is
       | replaced with Chinese dialogue. so, not very reliable in my
       | books.
        
       | modzu wrote:
       | image generation kind of reminds me of video games or any cgi in
       | general.. the progress is undeniable, and yet with every
       | milestone it seems the last gap to "photorealism" is infinitely
       | wide
        
       ___________________________________________________________________
       (page generated 2026-02-10 23:00 UTC)