Skip to main content
December 8, 2025
Question

Firefly Image Model v1 in Photoshop 27 generates low-resolution results – is this regional or global

  • December 8, 2025
  • 1 reply
  • 147 views

Title:
Firefly Image Model v1 in Photoshop 27 produces low-resolution and blurry results – is this a regional, account-type, or global issue?

Post:
Hello Adobe team and community members,

I would like to report a technical issue with Generative Fill in Photoshop 27 and ask for clarification from engineering and other users, to understand whether this is:

• a global Firefly problem
• a Photoshop 27.x bug
• a regional limitation (I am located in Serbia)
• an account-type limitation (VIP Marketplace license)
• or something specific to my installation

Here is the problem:

Even when I manually select *Firefly Image Model 1 (October 2023)* from the Model Selector, the generated areas have extremely low resolution and blurry texture. Clothing edges, fabric detail, skin texture, and small objects appear noticeably softer than the surrounding parts of the image.

This low-resolution behavior did **not** exist in earlier Photoshop versions, where Firefly v1 generated sharp and realistic results consistent with the original photo.

Right now, even when Model v1 is selected, the results look very similar to Model v3 — as if both models were using the same lower-resolution generative pipeline.

I am also aware that some Firefly and generative AI features are region-restricted in Serbia, so I would like to know:

• Is Firefly Model v1 inside Photoshop 27 technically identical to the original Firefly v1 that produced high-resolution outputs before the update?
• Or is the current version modified, simplified, or internally tied to Firefly v3 (or region-restricted) in a way that limits its output quality?

Before assuming this is a global issue, I would like to check whether other users — especially those outside Serbia — are experiencing the same degradation.

If anyone can confirm whether Model v1 still produces full-resolution, high-detail generations on their system, please share an example or a quick test.

Any technical clarification from Adobe engineers would be greatly appreciated, as this issue significantly affects professional workflows that rely on realistic generative detail.

Thank you.

1 reply

creative explorer
Community Expert
Community Expert
December 13, 2025

@dragoslav1960 from my undersatnding if you are creating a 300ppi image and use AI, it would process the AI for 300dpi; and the same goes if you are creating an image for the web at 72ppi, it would work and create an AI image at that ppi. Sometimes, I do need to apply a 'Unsharp Mask' or 'Sharpen More' depending on what I have. 

m
D Fosse
Community Expert
Community Expert
December 13, 2025

Generative AI is limited to 2000 x 2000 pixels to make it workable over internet bandwidth and the bandwidth of the servers. Any bigger and the whole thing would just choke. 

 

The ppi number does not come into it.