FLUX.2

“Flux 2 Klein or Qwen Image 2.1 might give you a less frustrating experience at comparable or better performance.”

“It's really fast on the OpenRouter API, but it's clearly worse than Flux 3's normal model.”

“it's clearly worse than Flux 3's normal model”

“Some of the stylized art I'm getting out of it at higher resolutions are really nice and I prefer them over other models.”

“its editing capabilities have been really solid, but I've been disappointed with its ability to do finer text details”

“The "print on a shirt" example looks so bad that I can't imagine a human looked at this and said”

“Text-to-image with strong prompt following and a native understanding of composition.”

“Love to see an AI lab outside of US/China releasing good models.”

“this is pretty impressive, well done, it seems to be filling decently a gap that everyone that has worked enough generating images with AI has faced”

“Dislike that it's yet another platform held back by arbitrary moderation.”

“The UX looks amazing and very steerable, congrats to the team for focusing on the interface.”

“It’s great to see this improve on the platform with declarative controls built into the API”

“eternally greatful these models are marked for open weight releases”

“it's going to require an absolute monster truck of a machine and still run significantly slower”

“If Flux3 is anything like Flux2, it's going to require an absolute monster truck of a machine”

“it produces incredibly consistent and stable results that work for my use case”

“It initially required significantly more VRAM and was much slower than alternatives released around the same time”

“maybe it's perhaps not totally in the right direction to only focus on visual stuff”

“All generations i have see looked blurry, lacking in fine details and missing the artistic touch”

“there's been a lot of positive movement in the open-weights space”

“Running locally via something like Comfy gets incredible results.”

“Writing an entire local generated asset pipeline using flux and hunyuan3D-2.1 was a really fun experience”

“Qwen-Image-2512 / Z-Image / Flux.2 absolutely crush SDXL if you're actually generating moderately complex scenes.”

“The technology has gotten good enough that the results are genuinely suitable for LinkedIn, email signatures, professional bios.”

“It's probably the best-in-class for sheer prompt adherence right now”

“It works, it's fast and the code quality is as high as I've ever seen a C code base.”

“I expected this C implementation to be notably faster, but my M3 Max (36GB) could barely make it past the first denoising step before OOMing”

“Z-Image is roughly as censored as Flux 2, from my very limited testing.”

“it's worse at all of those things other than perhaps looking like it was shot on a cell phone camera than Qwen Image and Flux 2”

“If Flux Base is easy to train and works as well with multiple LoRAs as SDXL, quite probably, yes”

“prompt adherence and scene/character consistency in edit mode is (way) ahead of Qwen-Edit-2509”

“Weak world knowledge, worse licensing, and it ruins the #1 benefit of a larger LLM backbone”

“Flux 2 is dead in the water even though the knowledge it has contained in the model is way, way higher”

“whereas BFL (makers of Flux 2) dedicated like a fith of their press release talking about how "safe" (read: censored and lobotomized) their model is”

“it's pretty incredible a tiny startup like BFL has become the primary alternative to Google”

“They should have creds for open weigths, and bein transparent of it not being open source though.”

“Those cases imply commercial workflows that are prohibited with the open-weights model without purchasing a license.”

“flux 1 dev one was never really great at artistic style, mostly locking you into a somewhat generic style”

“Great, especially that they still have an open-weight variant of this new model too.”

“No one else currently strikes the proper balance between following style / composition / text rendering quite as well”