Skip to content
ImagesFree

Image Editing Models Compared Against Original

A review post comparing one edit prompt across NanoBanana Flash, NanoBanana Pro, and GPT Image 2 against the original MidJourney render, noting loss of fine textures and degradation of composition and lighting, and calling significant edits with perfect preservation the weakest link in the AI filmmaking pipeline.

OpenAI

Source previews

These example images come from the original source. DROPIDEA has not independently tested these examples; results depend on the model and settings.

Ready to personalize

Your prompt

English

Replace the red variables with your information. Copy the full prompt and use it with your preferred agent.

TXT MD English version
The edit prompt I asked was very simple: "Replace the kid on the cliff with Kai (character sheet reference image) and keep same acting posture as he is looking towards the valley. Fix some of the small people scale who look bigger than their houses down in the valley." Here are the results. Image 1: Original MidJourney render. Image 2: NanoBanana Flash. Image 3: NanoBanana Pro. Same edit. Same prompt. Different model. 👉Flash is great for speed. But if you're preparing keyframes for Veo, Seedance, Kling, Higgsfield, or any cinematic workflow where every frame matters... ...don't use Flash. 👉The Pro model preserves noticeably more of the original composition, lighting, textures, and material detail. That said... Even Pro still degrades a surprising amount of what MidJourney created. 😅 - Fine textures disappear. - Materials become softer. - Small design decisions get simplified. - The image slowly drifts away from the original art direction. 👉GPT Image 2...well see for yourself...I'm not impressed. Right now, I honestly don't think there's a model that can make significant edits while perfectly preserving everything that made the original image great. It's currently the weakest link in the AI filmmaking pipeline. Hopefully not for long.
Preview results

Your prompt

The edit prompt I asked was very simple: "Replace the kid on the cliff with Kai (character sheet reference image) and keep same acting posture as he is looking towards the valley. Fix some of the small people scale who look bigger than their houses down in the valley." Here are the results. Image 1: Original MidJourney render. Image 2: NanoBanana Flash. Image 3: NanoBanana Pro. Same edit. Same prompt. Different model. 👉Flash is great for speed. But if you're preparing keyframes for Veo, Seedance, Kling, Higgsfield, or any cinematic workflow where every frame matters... ...don't use Flash. 👉The Pro model preserves noticeably more of the original composition, lighting, textures, and material detail. That said... Even Pro still degrades a surprising amount of what MidJourney created. 😅 - Fine textures disappear. - Materials become softer. - Small design decisions get simplified. - The image slowly drifts away from the original art direction. 👉GPT Image 2...well see for yourself...I'm not impressed. Right now, I honestly don't think there's a model that can make significant edits while perfectly preserving everything that made the original image great. It's currently the weakest link in the AI filmmaking pipeline. Hopefully not for long.

Variable guide

"Replace the kid on the cliff with Kai (character sheet reference image) and keep same acting posture as he is looking towards the valley. Fix some of the small people scale who look bigger than their houses down in the valley."

The complete quoted edit instruction at the top of the post; replaceable with another edit instruction while keeping quotation marks

Start of the English template, the text inside the quotation marks in the first line

Example value

"Replace the kid on the cliff with Kai (character sheet reference image)..."

What you get

  • A complete Arabic translation of a review post documenting a single-image edit experiment.
  • The original edit prompt preserved as a replaceable quoted block.
  • A direct comparison between NanoBanana Flash, NanoBanana Pro, and GPT Image 2.
  • A three-bullet list detailing degradation points (textures, materials, design, art direction drift).
  • An evaluative conclusion about the weakest link in the AI filmmaking pipeline.

Use cases

  • Sharing image-editing model comparisons on social media.
  • Guiding AI filmmakers when choosing a model for keyframe preparation.
  • Documenting current editing model limits in technical blogging.
  • A baseline for re-testing the same prompt on newly released models.

How to use

  • Read the quoted edit prompt at the top and decide whether to adapt it for your own test.
  • Upload the original MidJourney image and a character sheet reference image for the replacement character when replicating the experiment.
  • Run the prompt as-is, or replace the EDIT_PROMPT variable with your modified instruction while keeping the quotation marks.
  • Execute the same prompt across multiple models (Flash then Pro) using the same reference image for a fair comparison.
  • Compare the original and results side by side, focusing on texture, lighting, and composition.
  • Use the closing remarks to decide which model suits your workflow (speed vs cinematic quality).

Related prompts