120 lesson thumbnails: 0.5K Nano Banana 2.1 vs GPT low, cost on Sume

120 course lesson thumbnails cost $9.00 on Nano Banana 2.1 at 0.5K or $2.97 on GPT Image 2.5 low on Sume. Google direct would be $4.03 at 1K.

4 min readSume
All posts

The short answer

One hundred twenty lesson thumbnails cost $9.00 on Sume with Nano Banana 2.1 at 0.5K (120 x $0.075) and $2.97 with GPT Image 2.5 low at 1K (120 x $0.02475). Nano Banana 2.1 at 1K is $12.00.

Lesson thumbnails are small on screen, so the 0.5K tier is a reasonable fit, and the Google pricing page I read lists no 0.5K row, so Sume's $0.075 is the only 0.5K price in this comparison.

The arithmetic

A course with 12 modules and 10 lessons each has 120 lessons. If every thumbnail needs one retry, double the totals. The table shows a single pass.

Google lists $0.0336 for 1K, so 120 images direct would be $4.03 at its standard rate and $2.02 at its batch rate, shown below for comparison.

120 lesson thumbnails, as of 2026-10-08 (totals rounded to the cent; Google rows are references and not sold through Sume)
OptionCount x unit priceTotal
Nano Banana 2.1 at 0.5K120 x $0.075$9.00
Nano Banana 2.1 at 1K120 x $0.10$12.00
GPT Image 2.5 low (1K)120 x $0.02475$2.97
GPT Image 2.5 medium (2K)120 x $0.055625$6.68
Google direct, 1K standard (not on Sume)120 x $0.0336$4.03
Google direct, 1K batch (not on Sume)120 x $0.0168$2.02

How to run it

Use one style prompt for the whole course and change only the lesson subject. Consistent thumbnails help learners scan a long syllabus.

Add the lesson title as an overlay in your own tool rather than asking the model to draw the words, since long titles are where image models most often fail.

Retries and cost per lesson

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 lessons to give a per-lesson cost.

Treat the per-lesson 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 0.5K: $9.00 becomes $10.80 with 20 percent extra images, which is $0.09 per lesson (a 120-lesson job).
  • Nano Banana 2.1 at 1K: $12.00 becomes $14.40 with 20 percent extra images, which is $0.12 per lesson (a 120-lesson job).
  • GPT Image 2.5 low (1K): $2.97 becomes $3.56 with 20 percent extra images, which is $0.03 per lesson (a 120-lesson job).
  • GPT Image 2.5 medium (2K): $6.68 becomes $8.02 with 20 percent extra images, which is $0.07 per lesson (a 120-lesson job).

What the docs say to check

Read GET /v1/images/models before you pin a tier or ratio. A model accepts only the values its descriptors list, and the descriptors are the definitive source for what a call may send.

The endpoint pricing lines are what Sume charges your wallet and already include the Sume margin, so a call costs cost_usd times n. Sume bills image generation all or nothing: a generation that fails or is cancelled is not charged.

Run the first module as a pilot of 10 thumbnails. At $0.075 each that is $0.75 and tells you whether the style works before the other 110.

Every price in this post is a snapshot dated 2026-10-08. The Sume figures come from the public catalog and the Image API docs, the Google figures come from its Gemini API pricing page, and no figure here is a benchmark of quality or speed. Prices and model lists change, so confirm them on the day you spend money.

Sources

Related posts

More in Use cases

All Use cases posts

Written by Sume