gemini-3.1-flash-image deprecated: Nano Banana 2.1 on Sume
Google deprecated gemini-3.1-flash-image on Oct 6 when Nano Banana 2.1 went GA. Which ids Sume accepts, what it bills per image, and what Google lists.

Google's release notes dated October 6, 2026 say Nano Banana 2.1 is generally available and that gemini-3.1-flash-image is now deprecated. If your client still sends that id, plan the swap. On Sume the id to send is nano-banana-2.1 (or google/nano-banana-2.1), and the older Nano Banana 2 strings still run as 2.1.
Sume does not say it accepts gemini-3.1-flash-image. Look the id up in GET /v1/catalog rather than assuming, and expect 404 model_not_found for an id that is not listed.
Prices side by side
The two columns come from different rate cards and should not be compared as a markup. Sume's numbers are the catalog price per image at default options. Google's are its Standard tier from its pricing page, read 2026-10-08.
For 100 images at 1K, that is 100 x $0.10 = $10.00 on Sume and 100 x $0.0336 = $3.36 at Google's Standard rate. At 2K it is 100 x $0.15 = $15.00 against 100 x $0.0504 = $5.04. Google also lists a Batch tier at $0.0168, $0.0252 and $0.0567 per image for 1K, 2K and 4K.
| Size | Sume | Google Standard | Google Batch |
|---|---|---|---|
| 0.5K | $0.075 | not listed | not listed |
| 1K | $0.10 | $0.0336 | $0.0168 |
| 2K | $0.15 | $0.0504 | $0.0252 |
| 4K | $0.20 | $0.113 | $0.0567 |
Migration checklist
The code change is small, but the checks are not. Resolution names map directly: both sides list 1K, 2K and 4K. Output ids, aspect ratios and the way you receive the image differ, since Sume returns durable media.sume.com URLs from an async job.
- Replace the model string and resend a single request with
mode: "async", then poll the job. - Send an
Idempotency-Keyper order so a retry does not create a second paid job. - Read the result artifacts from the job and use the Sume media URL, not a provider URL.
- Run a 10-image sample and compare quality at 1K before moving a 1,000-image workload.
- Search your repos for the old id. The linked script post shows one way.
When Sume is the wrong choice
If you only need the lowest per-image price and already have a Google key, the Standard and Batch rates above are lower. Sume's reasons to switch are one key, one job model and one webhook scheme across image, video and audio, and the unified queue behavior in the Generation admission page. That trade is yours to make.
Budgeting the swap
Take your monthly image count by size. If you make 2,000 images at 1K and 500 at 2K, the Sume list cost is 2,000 x $0.10 + 500 x $0.15 = $200 + $75 = $275. At Google's Standard rate the same mix is 2,000 x $0.0336 + 500 x $0.0504 = $67.20 + $25.20 = $92.40. The Batch tier would be 2,000 x $0.0168 + 500 x $0.0252 = $33.60 + $12.60 = $46.20.
Those gaps are real, so decide what you are buying. Sume bills the job result and handles the queue, the retry key and the webhook; a direct Google integration does not. Run the arithmetic on your own mix before you decide, and keep the deprecation date on your calendar either way.
Sources
Related posts
More in Developers
- Gemini 3.7 Flash now routes to 3.8: where a media client pins ids
Google auto-routes gemini-3.7-flash to gemini-3.8-flash since Oct 8. Where Sume lets you pin a model, where it routes for you, and what the job records.
- Gemini Deep Research agent shuts down Oct 23: an async alternative
Google marked deep-research-pro-preview-12-2025 for shutdown on Oct 23, 2026. What an async Sume Agent Completion run does, and what it does not do like it.
- Gemini Omni Flash 1.1 prompt length on Sume: a 20,000-character check
The Sume catalog constraint for Gemini Omni Flash 1.1 is a prompt of at most 20,000 characters, 3 to 10 seconds. A short Python preflight check before you pay.
- Omni Flash defaults to 720p and 16:9: set both on every request
Google's Omni docs default to 720p and 16:9; Sume's Auto defaults to 720p and 8 seconds. Why to send resolution, ratio and duration every time, with prices.
Written by Sume