Virtual staging for 24 rooms: $3.60 on Nano Banana 2.1, $1.34 on GPT

The short answer

Staging 24 rooms with one edit each costs $3.60 on Sume with Nano Banana 2.1 at 2K (24 x $0.15) and $1.34 with GPT Image 2.5 medium at 2K (24 x $0.055625 = $1.335). GPT Image 2.5 is also the model with an optional mask_url, which limits an edit to the region you mark.

A mask matters for real estate because the walls, windows and floor must stay as photographed, and only the furniture should change.

The arithmetic

The mask is the reason to pick GPT Image 2.5 for staging. Nano Banana 2.1 does not list mask_url, so a request that sends it is rejected with 400 unsupported_parameter. You can still use Nano Banana 2.1 for the edit and describe the region in words.

The table prices one edit per room. A listing usually needs the same room staged in two styles, which doubles each total.

How to run it

Host the photo and the mask on public HTTPS URLs. The docs describe mask_url as a public HTTPS mask URL for ChatGPT Image 2.5 edits, and the reference rules reject localhost and private-network addresses.

Use aspect_ratio auto on edit calls so the output matches the source photo. If you omit the field, the result is not the same as auto.

Retries and cost per room

Some images will be rejected by eye, and rejected images are billed like any other, so budget a margin. The bullets add 20 percent to each Sume total in the table above (the totals multiplied by 1.2) and divide by the number of rooms to give a per-room cost.

Treat the per-room figure as the number to compare across models, because it does not change when the count changes. Re-read the public catalog before a large run, since prices can change and the table is a snapshot from 2026-10-08.

• Nano Banana 2.1 at 2K: $3.60 becomes $4.32 with 20 percent extra images, which is $0.18 per room (a 24-room job).
• Nano Banana 2.1 at 1K: $2.40 becomes $2.88 with 20 percent extra images, which is $0.12 per room (a 24-room job).
• GPT Image 2.5 medium (2K): $1.34 becomes $1.61 with 20 percent extra images, which is $0.07 per room (a 24-room job).
• GPT Image 2.5 low (1K): $0.59 becomes $0.71 with 20 percent extra images, which is $0.03 per room (a 24-room job).

What the docs say to check

mask_url and background (auto, transparent or opaque) are documented for ChatGPT Image 2.5. A request that sets a parameter the chosen model does not list is rejected with 400 unsupported_parameter, not silently dropped.

On GPT Image 2.5 the quality values are auto, low, medium, high, xhigh and max. If you omit quality the default is high, and auto reserves max, so pin the tier you budgeted for instead of leaving it to the default.

Disclose staged images as staged in the listing. The edit changes what a buyer will see in person.

The numbers above are list prices read on 2026-10-08, not measurements. Sume figures are the catalog prices that already include its margin, Google figures are from its own pricing page, and quality or speed differences are not covered. Re-read both pages before you commit a large budget, because either side can change its rates.