4 ms·

FLUX.2 [Klein]: Towards Interactive Visual Intelligence

227 points by GaggiX 9mo ago

codezero 9mo ago
I am amazed, though not entirely surprised, that these models keep getting smaller while the quality and effectiveness increases. z image turbo is wild, I'm looking forward to trying this one out. An older thread on this has a lot of comments: https://news.ycombinator.com/item?id=46046916 https://news.ycombinator.com/item?id=46046916
- roenxi 9mo ago
  There are probably some more subtle tipping points that small models hit too. One of the challenges of a 100GB model is that there is non-trivial difficulty in downloading and running the thing that a 4GB model doesn't face. At 4GB I think it might be reasonable to assume that most devs can just try it and see what it does.
- Auracle 9mo ago
  Quality is increasing, but these small models have very little knowledge compared to their big brothers (Qwen Image/Full size Flux 2). As in characters, artists, specific items, etc.
  - efskap 9mo ago
    I smell the bias-variance tradeoff. By underfitting more, they get closer to the degenerate case of a model that only knows one perfect photo.
  - vunderba 9mo ago
    Agreed - given what Tongyi-MAI Lab was able to accomplish with a 6b model - I would love to see what they could do with something larger. Somewhere in the range of 15-20b, between these smaller models (ZiT, Klein) and the significantly larger models (Flux.2 dev).
  - littlestymaar 9mo ago
    That's what LoRAs are for. And small models are also much easier to fine tune than large ones.
    - Auracle 9mo ago
      I hate that excuse. I want the model to know who the Paw Patrol is without either finding a lora (which probably won't exist because they're mostly porn) or needing to make a dataset, tag it, and then train it myself.
- aitchnyu 9mo ago
  Is there a theoritical minimum for params for a given output? I saw news about GPT 3.5, then Deepseek training models at a fraction of that cost, then laptops running a model that beats 3.5. When does it stop?
SV_BubbleTime 9mo ago
Flux2 Klein isn’t some generation leap or anything. It’s good, but let’s be honest, this is an ad. What will be really interesting to me is the release of Z-image, if that goes the way it’s looking, it’ll be natural language SDXL 2.0, which seems to be what people really want. Releasing the Turbo/Distilled/Finetune months ago was a genius move really. It hurt Flux and Qwen releases on a possible future implication alone. If this was intentional, I can’t think of the last time I saw such shrewd marketing.
- refulgentis 9mo ago
  I’m a bit confused, both you and another commenter mention something called Z-Image, presumably another Flux model? Your frame of it is speculative, i.e. it is forthcoming. Theirs is present tense. Could I trouble you to give us plebes some more context? :) ex. Parsed as is, and avoiding the general confusion if you’re unfamiliar, it is unclear how one can observe “the way it is looking”, especially if turbo was released months ago and there is some other model that is unreleased. Chose to bother you because the others comment was less focused on lab on lab strategy.
  - ollin 9mo ago
    Z-Image is another open-weight image-generation model by Alibaba [1]. Z-Image Turbo was released around the same time as (non-Klein) FLUX.2 and received generally warmer community response [2] since Z-image Turbo was faster, also high-quality, and reportedly better at generating NSFW material. The base (non-Turbo) version of Z-Image is not yet released. [1] https://tongyi-mai.github.io/Z-Image-blog/ https://tongyi-mai.github.io/Z-Image-blog/ [2] https://www.reddit.com/r/StableDiffusion/comments/1p9uu69/no_hard_feelings/ https://www.reddit.com/r/StableDiffusion/comments/1p9uu69/no...
    - refulgentis 9mo ago
      Ahh I see, and Klein is basically a response to Z-Image Turbo, i.e. another 4-8B sized model that fits comfortably on a consumer GPU. It’ll be interesting to see how the NSFW catering plays out for the Chinese labs. I was joking a couple months ago to someone that Seedream 4’s talents at undressing was an attempt to sow discord and it was interesting it flew under the radar. Post-Grok going full gooner pedo, I wonder if it Grok will take the heat alone moving forward.
psubocz 9mo ago
> FLUX.2 [klein] 4B The fastest variant in the Klein family. Built for interactive applications, real-time previews, and latency-critical production use cases. I wonder what kind of use cases could be "latency-critical production use cases"?
- drellybochelly 9mo ago
  Maybe fast image editing, since it supports that.
- satvikpendem 9mo ago
  Local models. I'm not gonna wait 10 min for one image on my computer like I did back in the Stable Diffusion days. And image editing in particular.
pavelstoev 9mo ago
If we think of GenAI models as a compression implementation. Generally, text compresses extremely well. Images and video do not. Yet state-of-the-art text-to-image and text-to-video models are often much smaller (in parameter count) than large language models like Llama-3. Maybe vision models are small because we’re not actually compressing very much of the visual world. The training data covers a narrow, human-biased manifold of common scenes, objects, and styles. The combinatorial space of visual reality remains largely unexplored. I am looking towards what else is out there outside of the human-biased manifold.
- murderfs 9mo ago
  > Generally, text compresses extremely well. Images and video do not. Is that actually true? I'm not sure it's fair to compare lossless compression ratios of text (abstract, noiseless) to images and video that innately have random sampling noise. If you look at humanly indistinguishable compression, I'd expect that you'd see far better compression ratios for lossy image and video compression than lossless text.
  - regularfry 9mo ago
    The comparison makes sense in what I am charitably assuming is the case the GP is referring to: we know how to build a tight embedding space from a text corpus, and get out outputs from it tolerably similar to the inputs for the purposes they're put to. That is lossy compression, just not in the sense anyone talking about conventional lossless text compression algorithms would use the words. I'm not sure we can say the same of image embeddings.
- CamperBob2 9mo ago
  Images and video compress vastly better than text. You're lucky to get 4:1 to 6:1 compression of text (1), while the best perceptual codecs for static images are typically visually lossless at 10:1 and still look great at 20:1 or higher. Video compression is much better still due to temporal coherence. 1: Although it looks like the current Hutter competition leader is closer to 9:1, which I didn't realize. Pretty awesome by historical standards.
dfajgljsldkjag 9mo ago
I appreciate that they released a smaller version that is actually open source. It creates a lot more opportunities when you do not need a massive budget just to run the software. The speed improvements look pretty significant as well.
airstrike 9mo ago
2026 will be the year of small/open models
pajtai 9mo ago
It cannot create an image of a pogo stick. I was trying to get it to create an image of a tiger jumping on a pogo stick, which is way beyond its capabilities, but it cannot create an image of a pogo stick in isolation.
- CamperBob2 9mo ago
  Those are both good benchmark prompts. Z-Image Turbo doesn't like them either: Tiger on pogo stick: https://i.imgur.com/lnGfbjy.jpeg https://i.imgur.com/lnGfbjy.jpeg Dunno what this is, but it's not a pogo stick: https://i.imgur.com/OmMiLzQ.jpeg https://i.imgur.com/OmMiLzQ.jpeg Nano Banana Pro FTW: https://i.imgur.com/6B7VBR9.jpeg https://i.imgur.com/6B7VBR9.jpeg
- vunderba 9mo ago
  It's a tough test for local models - (gpt-image and NB had zero problems) - the only one that came reasonably close was Qwen-Image Z-Image / Flux 2 / Hidream / Omnigen2 / Qwen Samples: https://imgur.com/a/tB6YUSu https://imgur.com/a/tB6YUSu This is where smaller models are just going to be more constrained and will require additional prompting to coax out the physical description of a "pogo stick". I had similar issues when generating Alexander the Great leading a charge on a hippity-hop / space hopper.
- mhl47 9mo ago
  You are right, just tried even with reference images it can't do it for me. Maybe with some good prompting. Because in theory I would say that knowledge is something that does not have to be baked in the model but could be added using reference images if the model is capable enough to reason about them.
- nomel 9mo ago
  When given an image of an empty wine glass, it can't fill it to the brim with wine. The pogo stick drawers and wine glass fillers can enjoy their job security for months to come!
  - downboots 9mo ago
vunderba 9mo ago
I haven’t gotten around to adding Klein to my GenAI Showdown site yet, but if it’s anything like Z-Image Turbo, it should perform extremely well. For reference, Z-Image Turbo scored 4 out of 15 points on GenAI Showdown. I’m aware that doesn’t sound like much, but given that one of the largest models, Flux.2 (32b), only managed to outscore ZiT (a 6b model) by a single point and is significantly heavier-weight, that’s still damn impressive. Local model comparisons only: https://genai-showdown.specr.net/?models=fd,hd,kd,qi,f2d,zt https://genai-showdown.specr.net/?models=fd,hd,kd,qi,f2d,zt
- Bombthecat 9mo ago
  Can you fix the information bubble on mobile please? When pressing one, it vanishes instantly...
  - vunderba 9mo ago
    Hey Bombthecat, sorry about that! I can't repro this issue on any of the devices I have (Android Pixel 7, an iPad, etc). If you get a chance, could you list your mobile device specs? That way I can at least try it on Browserstack and see if I can figure out a fix.
    - Bombthecat 9mo ago
      Samsung, brave browser Update: Huh, now it's working
    - kennyadam 9mo ago
      Yeah works fine for me on a Pixel 9.
- BoredPositron 9mo ago
  I think it shows problems with your tests tbh. The bigger models are way more capable than you make them out to be. They are also better in training and understanding of CGI render outputs as reference like normal maps or id-masks. Your testing suite is the perfect example that structured data implies false confidence. Pure t2i is not a good benchmark anymore.
tonyhart7 9mo ago
damn, they really counter attack after z-image release huh good competition breed innovation
Mashimo 9mo ago
Neat, I really enjoyed flux 1. Currently use z image turbo for messing around. I will wait for invoke to add flux2 klein.
- aeon_ai 9mo ago
  +1 - zit in invoke has been amazing to use
Nora23 9mo ago
How does this compare to GPT version in terms of interactive capabilities?
Yash16 9mo ago
Has anyone tried Flux 2 Klein? Lately, I’ve stopped chasing new models. I’m genuinely happy with Nano Banana Pro—the results are solid, so I’m building the entire app around just this one model. https://picxstudio.com https://picxstudio.com
wolvoleo 9mo ago
Ohh I thought this was about the tool that makes your screen more Orange at night. No longer necessary as every desktop has that built in. But it was also called F.Lux

227 points by GaggiX 9mo ago

I am amazed, though not entirely surprised, that these models keep getting smaller while the quality and effectiveness increases. z image turbo is wild, I'm looking forward to trying this one out. An older thread on this has a lot of comments: https://news.ycombinator.com/item?id=46046916 https://news.ycombinator.com/item?id=46046916

roenxi 9mo ago
There are probably some more subtle tipping points that small models hit too. One of the challenges of a 100GB model is that there is non-trivial difficulty in downloading and running the thing that a 4GB model doesn't face. At 4GB I think it might be reasonable to assume that most devs can just try it and see what it does.
Auracle 9mo ago
Quality is increasing, but these small models have very little knowledge compared to their big brothers (Qwen Image/Full size Flux 2). As in characters, artists, specific items, etc.
- efskap 9mo ago
  I smell the bias-variance tradeoff. By underfitting more, they get closer to the degenerate case of a model that only knows one perfect photo.
- vunderba 9mo ago
  Agreed - given what Tongyi-MAI Lab was able to accomplish with a 6b model - I would love to see what they could do with something larger. Somewhere in the range of 15-20b, between these smaller models (ZiT, Klein) and the significantly larger models (Flux.2 dev).
- littlestymaar 9mo ago
  That's what LoRAs are for. And small models are also much easier to fine tune than large ones.
  - Auracle 9mo ago
    I hate that excuse. I want the model to know who the Paw Patrol is without either finding a lora (which probably won't exist because they're mostly porn) or needing to make a dataset, tag it, and then train it myself.
aitchnyu 9mo ago
Is there a theoritical minimum for params for a given output? I saw news about GPT 3.5, then Deepseek training models at a fraction of that cost, then laptops running a model that beats 3.5. When does it stop?

SV_BubbleTime 9mo ago

Flux2 Klein isn’t some generation leap or anything. It’s good, but let’s be honest, this is an ad. What will be really interesting to me is the release of Z-image, if that goes the way it’s looking, it’ll be natural language SDXL 2.0, which seems to be what people really want. Releasing the Turbo/Distilled/Finetune months ago was a genius move really. It hurt Flux and Qwen releases on a possible future implication alone. If this was intentional, I can’t think of the last time I saw such shrewd marketing.

refulgentis 9mo ago
I’m a bit confused, both you and another commenter mention something called Z-Image, presumably another Flux model? Your frame of it is speculative, i.e. it is forthcoming. Theirs is present tense. Could I trouble you to give us plebes some more context? :) ex. Parsed as is, and avoiding the general confusion if you’re unfamiliar, it is unclear how one can observe “the way it is looking”, especially if turbo was released months ago and there is some other model that is unreleased. Chose to bother you because the others comment was less focused on lab on lab strategy.
- ollin 9mo ago
  Z-Image is another open-weight image-generation model by Alibaba [1]. Z-Image Turbo was released around the same time as (non-Klein) FLUX.2 and received generally warmer community response [2] since Z-image Turbo was faster, also high-quality, and reportedly better at generating NSFW material. The base (non-Turbo) version of Z-Image is not yet released. [1] https://tongyi-mai.github.io/Z-Image-blog/ https://tongyi-mai.github.io/Z-Image-blog/ [2] https://www.reddit.com/r/StableDiffusion/comments/1p9uu69/no_hard_feelings/ https://www.reddit.com/r/StableDiffusion/comments/1p9uu69/no...
  - refulgentis 9mo ago
    Ahh I see, and Klein is basically a response to Z-Image Turbo, i.e. another 4-8B sized model that fits comfortably on a consumer GPU. It’ll be interesting to see how the NSFW catering plays out for the Chinese labs. I was joking a couple months ago to someone that Seedream 4’s talents at undressing was an attempt to sow discord and it was interesting it flew under the radar. Post-Grok going full gooner pedo, I wonder if it Grok will take the heat alone moving forward.

psubocz 9mo ago

> FLUX.2 [klein] 4B The fastest variant in the Klein family. Built for interactive applications, real-time previews, and latency-critical production use cases. I wonder what kind of use cases could be "latency-critical production use cases"?

drellybochelly 9mo ago
Maybe fast image editing, since it supports that.
satvikpendem 9mo ago
Local models. I'm not gonna wait 10 min for one image on my computer like I did back in the Stable Diffusion days. And image editing in particular.

pavelstoev 9mo ago

If we think of GenAI models as a compression implementation. Generally, text compresses extremely well. Images and video do not. Yet state-of-the-art text-to-image and text-to-video models are often much smaller (in parameter count) than large language models like Llama-3. Maybe vision models are small because we’re not actually compressing very much of the visual world. The training data covers a narrow, human-biased manifold of common scenes, objects, and styles. The combinatorial space of visual reality remains largely unexplored. I am looking towards what else is out there outside of the human-biased manifold.

murderfs 9mo ago
> Generally, text compresses extremely well. Images and video do not. Is that actually true? I'm not sure it's fair to compare lossless compression ratios of text (abstract, noiseless) to images and video that innately have random sampling noise. If you look at humanly indistinguishable compression, I'd expect that you'd see far better compression ratios for lossy image and video compression than lossless text.
- regularfry 9mo ago
  The comparison makes sense in what I am charitably assuming is the case the GP is referring to: we know how to build a tight embedding space from a text corpus, and get out outputs from it tolerably similar to the inputs for the purposes they're put to. That is lossy compression, just not in the sense anyone talking about conventional lossless text compression algorithms would use the words. I'm not sure we can say the same of image embeddings.
CamperBob2 9mo ago
Images and video compress vastly better than text. You're lucky to get 4:1 to 6:1 compression of text (1), while the best perceptual codecs for static images are typically visually lossless at 10:1 and still look great at 20:1 or higher. Video compression is much better still due to temporal coherence. 1: Although it looks like the current Hutter competition leader is closer to 9:1, which I didn't realize. Pretty awesome by historical standards.

dfajgljsldkjag 9mo ago

I appreciate that they released a smaller version that is actually open source. It creates a lot more opportunities when you do not need a massive budget just to run the software. The speed improvements look pretty significant as well.

airstrike 9mo ago

2026 will be the year of small/open models

pajtai 9mo ago

It cannot create an image of a pogo stick. I was trying to get it to create an image of a tiger jumping on a pogo stick, which is way beyond its capabilities, but it cannot create an image of a pogo stick in isolation.

CamperBob2 9mo ago
Those are both good benchmark prompts. Z-Image Turbo doesn't like them either: Tiger on pogo stick: https://i.imgur.com/lnGfbjy.jpeg https://i.imgur.com/lnGfbjy.jpeg Dunno what this is, but it's not a pogo stick: https://i.imgur.com/OmMiLzQ.jpeg https://i.imgur.com/OmMiLzQ.jpeg Nano Banana Pro FTW: https://i.imgur.com/6B7VBR9.jpeg https://i.imgur.com/6B7VBR9.jpeg
vunderba 9mo ago
It's a tough test for local models - (gpt-image and NB had zero problems) - the only one that came reasonably close was Qwen-Image Z-Image / Flux 2 / Hidream / Omnigen2 / Qwen Samples: https://imgur.com/a/tB6YUSu https://imgur.com/a/tB6YUSu This is where smaller models are just going to be more constrained and will require additional prompting to coax out the physical description of a "pogo stick". I had similar issues when generating Alexander the Great leading a charge on a hippity-hop / space hopper.
mhl47 9mo ago
You are right, just tried even with reference images it can't do it for me. Maybe with some good prompting. Because in theory I would say that knowledge is something that does not have to be baked in the model but could be added using reference images if the model is capable enough to reason about them.
nomel 9mo ago
When given an image of an empty wine glass, it can't fill it to the brim with wine. The pogo stick drawers and wine glass fillers can enjoy their job security for months to come!
- downboots 9mo ago

vunderba 9mo ago

I haven’t gotten around to adding Klein to my GenAI Showdown site yet, but if it’s anything like Z-Image Turbo, it should perform extremely well. For reference, Z-Image Turbo scored 4 out of 15 points on GenAI Showdown. I’m aware that doesn’t sound like much, but given that one of the largest models, Flux.2 (32b), only managed to outscore ZiT (a 6b model) by a single point and is significantly heavier-weight, that’s still damn impressive. Local model comparisons only: https://genai-showdown.specr.net/?models=fd,hd,kd,qi,f2d,zt https://genai-showdown.specr.net/?models=fd,hd,kd,qi,f2d,zt

Bombthecat 9mo ago
Can you fix the information bubble on mobile please? When pressing one, it vanishes instantly...
- vunderba 9mo ago
  Hey Bombthecat, sorry about that! I can't repro this issue on any of the devices I have (Android Pixel 7, an iPad, etc). If you get a chance, could you list your mobile device specs? That way I can at least try it on Browserstack and see if I can figure out a fix.
  - Bombthecat 9mo ago
    Samsung, brave browser Update: Huh, now it's working
  - kennyadam 9mo ago
    Yeah works fine for me on a Pixel 9.
BoredPositron 9mo ago
I think it shows problems with your tests tbh. The bigger models are way more capable than you make them out to be. They are also better in training and understanding of CGI render outputs as reference like normal maps or id-masks. Your testing suite is the perfect example that structured data implies false confidence. Pure t2i is not a good benchmark anymore.

tonyhart7 9mo ago

damn, they really counter attack after z-image release huh good competition breed innovation

Mashimo 9mo ago

Neat, I really enjoyed flux 1. Currently use z image turbo for messing around. I will wait for invoke to add flux2 klein.

aeon_ai 9mo ago
+1 - zit in invoke has been amazing to use

Nora23 9mo ago

How does this compare to GPT version in terms of interactive capabilities?

Yash16 9mo ago

Has anyone tried Flux 2 Klein? Lately, I’ve stopped chasing new models. I’m genuinely happy with Nano Banana Pro—the results are solid, so I’m building the entire app around just this one model. https://picxstudio.com https://picxstudio.com

wolvoleo 9mo ago

Ohh I thought this was about the tool that makes your screen more Orange at night. No longer necessary as every desktop has that built in. But it was also called F.Lux