Nano Banana 3: does it exist? Current ids
Google autocomplete suggests a Nano Banana 3, but suggestions are not releases. What autocomplete lists and which Nano Banana id Sume accepts.

Autocomplete showing "nano banana 3" proves people search for it, not that a model with that name exists. On the read date, Google's suggestions for "nano banana" included 3 alongside pro, 2, gemini, ai, 2 pro, free, 2 lite and api. Do not put a guessed id into an API call; read the model list of the platform you call.
What the suggestion list shows
The table is the autocomplete response for the query "nano banana" on the read date. Order and wording are Google's and change over time, so this is evidence of interest and of naming patterns only.
| Suggestion | What it indicates |
|---|---|
| nano banana pro | A Pro variant is searched |
| nano banana 2 | A second version is searched |
| nano banana 2 pro | Version and Pro combined |
| nano banana 2 lite | A lite tier is searched |
| nano banana 3 | Searched; not a confirmation of a release |
| nano banana api | People look for API access |
Which id Sume accepts
The Sume Image API page lists legacy bare ids, naming gpt-image-2 and nano-banana-2 as examples, and says they are accepted as aliases for their org/slug equivalents. That is the only Nano Banana id the docs name.
The docs do not name a Nano Banana 3 or other variants, so a request for one would not be something to rely on. The reliable check is the catalog: GET /v1/images/models lists the ids and pricing the endpoint accepts. A model that is not in that response is not available to you.
curl "https://api.sume.com/v1/images/models" \
-H "Authorization: Bearer $SUME_API_KEY"A rule for future model names
When a new name appears in search, check three things before changing code: the vendor's own announcement, the model list of your API provider, and the exact id string the list uses. Pin the id from the list, not from a headline.
If you want Sume to choose, sume/auto is the Sume-only model value for images. Sume picks the family and never discloses which one ran, so use a pinned id when you need repeatable output.
Sources
Related posts
More in Models
- Lip sync with a hand over the mouth: sync-3 vs Fabric on Sume
Sync Labs says sync-3 handles obstructions on faces. Sume's lip-sync routes start from a still plus audio. What each takes as input and what is promised.
- Luma's 2026 timeline: Ray3.14, Ray3.2, Scenes and Variants
Luma shipped Ray3.14 in January, Ray3.2 in June, Scenes in August and Variants on Oct 1, 2026. What each added, and why to pin model ids.
- Lyria 3.5 blocks artist-voice prompts: how to write briefs that pass
Google's Lyria 3.5 docs note that prompts asking for specific artist voices are blocked. Describe the sound instead, then run it through the Sume Music Router.
- MAI-Voice-2.1 emotion control vs Sume's emotion field
Microsoft lists emotion control on MAI-Voice-2.1. Sume's TTS takes a free-text emotion string, speed and volume. What each gives you, and how to test it.
Written by Sume