APIMaster.ai

Nano Banana 2.5 API: Status, Model ID and Current Pricing

Nano Banana 2.5 is a useful search label, but it is not a separately confirmed public API ID. See its relationship to Gemini 2.5 Flash Image, current availability and APIMaster prices.

Nano Banana 2.5Nano Banana APIGemini 2.5 Flash ImageVertex AI image APIAI image generationAPIMaster

Published 2026-10-06

Quick Answer

Nano Banana 2.5 is not a dead model, but nano-banana-2.5 is not a confirmed public Google API ID. The name has appeared in Flow and community reports about Vertex testing, while Google’s published API naming maps the first Nano Banana generation to gemini-2.5-flash-image. That confirmed model still has an active APIMaster route in the latest check, even though Google has marked it deprecated and published a future Vertex retirement date.

On October 6, 2026, APIMaster’s live catalogue returned one active route for gemini-2.5-flash-image at from $0.03744 per 1K image. The current forward-looking sibling, Nano Banana 2 / gemini-3.1-flash-image, had three active routes at from $0.0469 per 1K image. Use the marketplace card as the source of truth for today’s availability and price.

Model ID Official/reference price Current APIMaster price Active routes Discount Model card
gemini-2.5-flash-image — current 2.5-era API $0.039 per 1K image From $0.03744 per 1K image 1 Up to ~4% off View routes
gemini-3.1-flash-image — Nano Banana 2 $0.067 per 1K image From $0.0469 per 1K image 3 Up to ~30% off View routes

The table reports the lowest active-route price, not a fixed quote for every channel. Google has not published a separate nano-banana-2.5 API row, so this article distinguishes the searchable label from the confirmed API model.

GPT Trial(GPT-6 Astra, GPT-6 Sol, GPT-5.6, and GPT-5.5 available)

Sign up and claim a $20 GPT trial

Claim now

What is Nano Banana 2.5?

The phrase currently combines three different layers that should not be conflated:

  1. A product label. Flow and other first-party surfaces have shown “Nano Banana 2.5 Flash” or related names during rollout tracking.
  2. A confirmed API model. Google’s public API documentation identifies the original Nano Banana image model as gemini-2.5-flash-image.
  3. Internal or partner identifiers. Community reports have mentioned names such as nano-banana-2.5 and internal image-generator strings, but these are not official public request IDs unless Google documents them.

That is why a search for “Nano Banana 2.5 API” can find real implementation clues without yielding a stable endpoint named nano-banana-2.5. The most accurate statement is that the 2.5-era capability is still accessible through the confirmed Gemini model where a route is available, while the branded alias itself remains unconfirmed.

Is Gemini 2.5 Flash Image still usable?

Yes, where a live route accepts it. APIMaster had one status=1 route on the verification date, so it would be inaccurate to tell developers that the model is already unusable. At the same time, Google has marked Gemini 2.5 Flash Image deprecated and lists March 15, 2027 as the Vertex retirement date. Deprecation means you should plan a migration; it does not erase an active route today.

For new production work, prefer Nano Banana 2 / gemini-3.1-flash-image unless you have a compatibility requirement for the 2.5-era behavior. For an existing project, pin gemini-2.5-flash-image, monitor errors and latency, and test the migration target before the retirement date.

Nano Banana 2.5 vs Nano Banana 2

The confirmed comparison is between API IDs, not between two aliases:

  • 2.5-era API: gemini-2.5-flash-image, one active APIMaster route in the October 6 check, and a deprecation notice from Google.
  • Nano Banana 2: gemini-3.1-flash-image, three active APIMaster routes in the same check and Google’s current public Nano Banana 2 mapping.
  • nano-banana-2.5: no separate official Google API page or public APIMaster route is claimed by this article.

The right choice depends on whether you are preserving compatibility or starting fresh. The 2.5 route is a practical bridge while it remains active; the 3.1 route is the better default for a new integration because its public naming and lifecycle position are clearer.

Current APIMaster pricing

The live price is a route observation, not a permanent list price. The Gemini 2.5 Flash Image card shows the current route and starts at $0.03744 per 1K image in the check used for this article. The Nano Banana 2 card starts at $0.0469 per 1K image across three active routes. Both prices are below the cited official reference prices, but the route list and discount can change.

APIMaster’s one-key setup lets you test both confirmed IDs without changing providers. That is useful for a migration: send the same prompts and reference images to both routes, compare usable output and latency, then choose based on the workload rather than the marketing label.

How to call the 2.5-era route

Create an APIMaster account, create a key in the console, and use the confirmed API ID rather than the unverified alias:

curl "https://apimaster.ai/v1/images/generations" \
  -H "Authorization: Bearer $APIMASTER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gemini-2.5-flash-image",
    "prompt": "A warm editorial product photograph of a glass teapot beside a lemon tree.",
    "size": "1:1",
    "resolution": "1K",
    "n": 1
  }'

If you are starting a new project, change only the model value to gemini-3.1-flash-image and verify its live card first. The endpoint and APIMaster key remain the same.

Migration checklist

  1. Separate label from ID. Store nano-banana-2.5 as an optional display label, never as an assumed request value.
  2. Pin the confirmed model. Use gemini-2.5-flash-image for compatibility or gemini-3.1-flash-image for a new default.
  3. Test the replacement now. Compare image quality, editing behavior, resolution, latency and cost on representative prompts.
  4. Monitor lifecycle dates. Keep the Google deprecation and March 15, 2027 Vertex retirement date in your migration calendar.
  5. Check the live route. APIMaster prices and availability can change, so re-open the model card before a production rollout.

FAQ

Is Nano Banana 2.5 an official API model?

Google has not published nano-banana-2.5 as a separate public Gemini API model ID in the official documentation checked for this article. The confirmed 2.5-era API ID is gemini-2.5-flash-image.

Can I still use Gemini 2.5 Flash Image?

Yes, when a live route accepts it. APIMaster had one active route on October 6, 2026. Google has deprecated the model and lists March 15, 2027 as the Vertex retirement date, so new projects should plan a migration rather than treat it as a long-term default.

What is the current Nano Banana alternative?

Nano Banana 2 maps to gemini-3.1-flash-image. It had three active APIMaster routes at from $0.0469 per 1K image in the October 6 check. See the live model card for current prices.

What is the cheapest 2.5-era route on APIMaster?

The gemini-2.5-flash-image card started at $0.03744 per 1K image across one active route on October 6, 2026, versus a $0.039 reference price. The route can change, so use the card rather than treating this as a guaranteed quote.

Is Nano Banana 2.5 the same as Nano Banana 2.1?

The available clues suggest a possible naming evolution, but Google has not officially documented the relationship. This article keeps both search terms separate and uses confirmed API IDs when discussing actual calls.

Should I build on nano-banana-2.5?

No. Build on gemini-2.5-flash-image if you need the confirmed 2.5-era route, or use gemini-3.1-flash-image for a new integration. Keep any product label in metadata, not in the request body.

Sources and verification date

Use a live image API while you migrate

Nano Banana 2.5 remains a useful search term and the confirmed 2.5-era route is still usable where available, but the alias is not a stable API contract. For a new integration, start with Nano Banana 2 / gemini-3.1-flash-image from $0.0469 per 1K image on the checked APIMaster routes. Register for APIMaster, create a key in the console, point your client at https://apimaster.ai/v1, and compare both live model cards before committing to a long migration.