Image to Image AI: Online Photo Editing Guide

Image to image AI uses a reference picture to guide an edit. Eight models in the Image studio accept a picture as input. The cheapest render costs 10 credits and the dearest costs 55. The 55-credit option is GPT Image 1.5 with its quality dial turned up, multiplying that model's own base price by 5.5; price alone does not establish that it is the best choice for your edit.

Credit figures reflect the SynthPulse configuration reviewed on October 23, 2026, not official vendor subscription or API prices. Check the live generation quote before submitting.

Compare the output size alongside the quote. Seedream 5.0 Lite and 4.5 include 2K at a flat rate; Seedream 5.0 Pro charges more for its 2K option. The table below separates those costs from prompt choices.

Image to image AI vs text to image

Text to image starts from nothing. Your words are the whole input and the model invents every pixel, which means each attempt re-rolls the composition, the pose, the framing, the lot. Get a shot you like on attempt four and attempt five will be a different shot.

Image to image AI starts from a picture you already have. The model reads that image, then applies the change you described on top of it. The reference can guide the room, product, face or layout, but preservation is a request, not a guarantee.

Choose the mode by your starting material:

• Nothing exists yet and you are describing it into being. Text to image.
• Something exists and needs to change. Image to image. Repaint a photo as an illustration, swap a background, drop a real product into a new setting, clean up an artefact, push forty product shots into one consistent look.

The mechanics are otherwise identical. Pick a model, write a prompt, press generate. For a reference edit, attach the picture first.

Image to image AI pricing: eight reference-compatible models

Cheapest first, in credits per image:

• Nano Banana — 10. One output size only.
• GPT Image 1.5 — 10 at medium quality, 55 at high. Square output only.
• Seedream 5.0 Lite — 14, flat. Renders at 2K.
• GPT Image 2 — 15 at 1K, 26 at 2K.
• Seedream 4.5 — 17, flat. Renders at 2K.
• Seedream 5.0 Pro — 18 at 1K, 36 at 2K.
• Nano Banana 2 — 20 at 1K, 30 at 2K.
• Nano Banana Pro — 45, and 2K costs the same 45.

Quantity multiplies all of that. You can ask for 1, 2 or 4 renders in one submission and every one of them is charged at full price; there is no batch rate. The number on the generate button is computed by the same code that actually deducts from your balance, so what you see there is what you pay.

Compare output size at the same budget

At 2K, the listed costs range from 14 to 45 credits. Equal dimensions do not mean equal results.

Seedream 5.0 Lite gives you a 2K image for 14 credits. It is flat-priced: there is no upgrade to buy, because its base tier is already the big one. Seedream 4.5 works the same way at 17.

Seedream 5.0 Pro charges 18 for 1K and 36 for 2K. So the Pro model's 2K render costs about two and a half times the Lite model's 2K render. Both produce 2K files. That comparison says nothing about which will preserve your product or handle a difficult edit better. If you're comparing them, use the same source and change request, then inspect the results.

Nano Banana Pro costs 45 credits at either size. Choose 2K if you need the larger file; 1K is also available at the same quote. More pixels do not guarantee a more accurate edit.

GPT Image 1.5's high tier is the most expensive option on this page. 10 credits to 55 is a 5.5x multiplier, which lands it above Nano Banana Pro's flat rate. OpenAI documents improved text rendering for the model, but that does not prove the high tier beats other models for your text-heavy edit. Try the cheaper tier first rather than paying for high by habit.

Check shape and size limits before writing a brief. GPT Image 1.5 only renders square, so every other aspect ratio in the picker disappears when you select it. Nano Banana has a single output size and no upgrade path at all.

Upload a reference and make one edit

1. Attach the reference. JPEG, PNG, WebP, GIF, AVIF and HEIC/HEIF are accepted, up to 3 MB per file. SVG is refused outright. If your source is a 12 MB export, downscale it first; the models work from a resized copy anyway.
2. Pick the model before you write the prompt. The aspect ratio and size controls change depending on what you selected, and so does the price on the button.
3. Describe the change. Add the details you need preserved; the examples below show both parts.
4. Set the output size. On Nano Banana Pro, 2K costs no extra credits. On Seedream 5.0 Pro, decide whether you actually need the 36-credit version.
5. Start with one image. Review it against the source before paying for more. Use a larger quantity only if you need alternatives.

Image to image AI prompts: name the edit and what stays

A long description of a new scene can obscure the edit you want. Start with the change, then name the features to preserve. You can repeat an important source detail without redescribing the entire photo.

Compare:

Bad: A red vintage car on a coastal road at sunset, shot on film, cinematic, highly detailed.

Good: Repaint this as a 1970s gouache travel poster. Flat colour blocks, visible paper grain, no photographic texture. Keep the car's position, angle and proportions exactly as they are.

Try these four edit patterns. They are suggested instructions, not benchmarked outputs:

Restyle

Convert this photograph into a pen-and-ink illustration with cross-hatched shadows. Line weight heavier on the foreground. Keep the framing and the subject's pose unchanged. White background where the sky is.

Background swap

Replace the background behind the subject with an overcast beach at low tide. Match the existing lighting direction, which comes from the upper left. Do not alter the subject, their clothing, or the edges of their hair.

Product placement

Place this bottle on a wet slate surface with shallow standing water under it. Single soft light from the right, one reflection. Keep the label text, the cap and the bottle's proportions identical to the reference.

Repair

Remove the power lines crossing the upper third of the image. Reconstruct the sky behind them to match the existing gradient and cloud texture. Change nothing else.

Each prompt names what should stay unchanged. Check those details against the source after generation; writing them down does not lock the pixels.

AI color changer: edit a product or clothing color

An AI color changer prompt is useful when you want a new color concept while keeping the source composition. On SynthPulse this uses the ordinary image-to-image editor. There is no separate region-selection or exact color-replacement tool.

Open the image studio with a product recolor prompt, attach your reference photo, then choose a compatible image-to-image model. The link fills the prompt; it does not attach a picture or start a paid generation.

Change product color with AI

Use a source with clear object edges and few reflections. A plain mug is easier to inspect than a transparent bottle with a printed label.

Change only the ceramic mug body in the reference photo to muted sage green. Preserve the handle shape, rim, position, background, light direction and contact shadow. Keep the same matte material. No added objects or text.

Compare the handle, rim and shadow with the source before accepting the color. If the model changes the shape, simplify the edit request. A good-looking variation is still unsuitable for a listing if it changes the actual product.

AI clothes color changer prompt

Change the jacket fabric to dark navy. Keep the person's face, pose, hair, buttons, seams and fabric folds as close to the reference as possible. Keep the trousers and background unchanged. Preserve the direction of the existing light.

For clothing, check edges beside hair and hands, the stitching and any logo. Describing navy is a creative direction, not a calibrated dye swatch. For an exact brand color or a production proof, use a selection-based image editor and a color-managed workflow.

What does an AI color change cost?

There is no separate recoloring fee: the selected model charges for an image generation. Check the quote after attaching the reference and choosing the output options. Start with one image, compare it with the source, and change one instruction per attempt.

The recolor prompts are suggested edits, not verified before-and-after examples. Related tasks use the same studio: generate a texture or plan a wedding portrait from a couple photo.

Combine more than one reference

Six of the eight models take up to 10 reference images in a single request. GPT Image 1.5 and Nano Banana take up to 16.

That is what makes combination prompts work: hand the model a subject in one picture and a jacket, a room or a colour palette in another, then say how the two should meet. Name the references by upload order and give each one a role.

Put the person from the first image into the room from the second. Keep their face, hair and clothing exactly as in the first image. Match the second image's lighting: warm lamp light from the right, deep shadow on the left wall.

“Combine these” leaves the relationship open. Specify which source supplies the person, which supplies the room and which details should remain.

When image to image AI is the wrong tool

Image to image AI is not a precision editor. If you need one pixel-exact region changed and everything else byte-identical, use an editor. The model re-renders the whole frame every time, so small unrequested drift in skin tone, grain and edges is normal rather than a bug.

Check object counts carefully. A request to remove one of five objects can also alter the remaining objects. Describe the state you want instead of the operation you want performed.

If the reference is a low-resolution crop, the model has less detail to preserve, and a 2K output is not proof that the missing detail was recovered accurately. Start from the best source you have and inspect the result closely.

Checks before generating

Before the first attempt, compare your source with the request:

• Can you see the whole feature you want to preserve, including its edges?
• Is the requested change distinguishable from the details that should stay?
• Does the selected model support your reference count, output shape and size?
• Is the quantity set to one, with enough credits for the displayed quote?

Afterward, inspect the changed area and the supposedly untouched areas. A new background can look convincing while a face, seam or label has drifted. Keep the original beside the result until you've checked both.

---

Model pages with the full picture on each family: Seedream, GPT Image, Nano Banana. Everything the platform sells, in one list: all models. Prompt-level detail on the newest of them: how to use Nano Banana 2 and Nano Banana 2 prompts. Credit packs and plans: pricing. To try an edit, open the Image studio.