Nano Banana 2 thumbnail tier: Google 0.5K to 4K vs Sume billed

Google's 2026 Nano Banana 2 per-image prices from 0.5K to 4K next to what Sume bills, to pick a thumbnail tier by cost rather than habit.

4 min readSume
All posts

A thumbnail rarely needs 4K. Nano Banana 2 prices by resolution tier, so the tier you pick is the largest lever on cost. Google's Gemini API pricing page lists Standard per-image prices for gemini-3.1-flash-image: $0.045 at 0.5K, $0.067 at 1K, $0.101 at 2K and $0.151 at 4K, from an output-token rate of $60 per million image tokens.

Sume's catalog lists the same model as google/nano-banana-2 with tiers 512 (alias 0.5K), 1K, 2K and 4K. Sume bills the fal list price times 1.25 (the SUME_STANDARD_MARGIN_RATIO documented in the repo), so the figures differ from Google's direct price. The list prices in the Sume pricing code are $0.06, $0.08, $0.12 and $0.16.

The ladder

Nano Banana 2 per image, USD (read 2026-10-05)
TierGoogle direct listSume list basisSume billed (x1.25)
0.5K0.0450.060.075
1K0.0670.080.10
2K0.1010.120.15
4K0.1510.160.20

Choosing a tier for a thumbnail

Google's page maps 1K to 1024 by 1024 pixels and 2K to 2048 by 2048. A card thumbnail shown at a few hundred pixels needs neither the 4K tier nor often the 2K tier. A sensible plan is to generate candidates at 1K, then rerun only the winner at 2K if you need a hero image.

  • Ten candidates at 1K on Sume: $1.00 billed.
  • Ten candidates at 2K on Sume: $1.50 billed.
  • Ten candidates at 4K on Sume: $2.00 billed.
curl -X POST https://api.sume.com/v1/images \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"google/nano-banana-2","prompt":"Bold product thumbnail, large empty left third","resolution":"1K","aspect_ratio":"16:9","n":4}'

Where the gap comes from

Going direct to Google is cheaper per image at every tier in the table. What the Sume price buys is one key, one wallet and the same request shape across 19 image ids, which matters when a thumbnail job falls back to another model. Compare on your own volume before you decide.

How this was checked

Vendor facts come from the pages listed in the sources, read on 2026-10-05. Sume facts come from the Image API docs and the catalog code on main on the same date. Catalogs and limits change, so read the descriptors from GET /v1/images/models before you pin a number in production code.

Sources

Related posts

More in Pricing

All Pricing posts

Written by Sume