Gemini Flash Image is the fast tier of Google's Gemini image line — the model you may know as Nano Banana 2. This page is about picking the right member of that line; if you want the version-by-version specs, the Nano Banana 2 page has them.
Flash is the fast middle of Google's image range, with a Pro tier above and a Lite tier below.
The tier where combining several photos into one image becomes possible.
Examples made on Kavel with this model.
AI-generated reference frames chosen to show what a fast tier is usually asked for — sets where the same treatment repeats across many panels, one subject rendered several ways, and grids of small frames. They illustrate the territory rather than being outputs captured from this specific model.
Gemini Flash Image is how Google's image model is named inside the Gemini product line, and it is the same model the internet knows as Nano Banana 2. Two names for one thing is confusing enough on its own, but the more useful question is which member of the line you want, because they are genuinely different tools rather than speed settings.
Flash is the fast middle: it accepts up to 14 reference photos in a single run, which is what makes composites and style transfer possible, it takes prompts up to 20,000 characters, and it outputs at 1K, 2K or 4K with the cost rising as you go up. Above it sits the Pro tier, which is the one to reach for when text inside the image has to stay legible or fine detail has to survive enlargement.
Below it sits the Lite tier, which takes a single source photo at a flat cost. On Kavel this runs through a managed provider channel with the credit estimate shown before you commit.
Creative engine
Lite for one photo at a flat cost, Flash for composites and resolution choice, Pro for type and detail.
Gemini Flash Image is the product-line name; Nano Banana 2 is the name it became known by.
Up to 14 reference photos in one run is the capability that separates Flash from the Lite tier.
Move up when in-image type or fine detail is the point of the picture.
The honest answer is that the tiers do not differ by quality alone, so picking by 'the best one' wastes money in one direction and disappoints in the other. Flash is right when you need several reference photos combined, when you want a say in output resolution, and when the run count is high enough that per-run cost is part of the decision. Pro is right when the image contains type that has to be readable or detail that has to survive being printed or enlarged. Lite is right when the job is one photo, one change, and you would rather not think about resolution pricing at all.
Up to 14 reference photos in one run — the thing the Lite tier cannot do at all.
In-image text and fine detail are where the top tier earns its higher cost.
One source photo at one cost regardless of resolution, which keeps quick edits simple.
The capabilities that define this position in the line, rather than a list every tier shares.
Combine up to 14 reference photos, moving elements and style between them.
1K, 2K or 4K, with the credit cost rising as the resolution does.
Up to 20,000 characters, which is enough to specify a scene in real detail.
No reference material required when you want to start from a description.
The generation parameters for this tier. It is the same model as Nano Banana 2, so these figures match that page.
The tier decision comes first, because it changes what inputs you can give.
Several source photos, or readable type, or neither — that answer picks the tier for you.
Write the scene, or upload up to 14 reference photos to combine.
1K for volume and iteration, 4K when the image is the finished artefact.
The estimate reflects the resolution you picked. Confirm it and run.
Jobs where the combination of composite input and a resolution choice is what makes them practical.
A run of panels or cards that must read as one family, generated in volume at a lower resolution.
Several source photos combined into one composed frame rather than shot together.
Moving the look of one reference onto the content of another, in a single run.
Many cheap 1K passes to settle the composition, then one 4K run for the version you keep.
One credit pool covers Nano Banana images and Seedance and Veo video. The cost shows before every run, and failed jobs are not charged. Use a subscription for ongoing work, or a one-time pack when you just need to top up.
What's included
Secure checkout by Stripe. Card details never touch our servers.
What's included
Secure checkout by Stripe. Card details never touch our servers.
What's included
Secure checkout by Stripe. Card details never touch our servers.
What's included
Secure checkout by Stripe. Card details never touch our servers.
One-time top-ups — buy extra credits any time you run low.
What's included
Secure checkout by Stripe. Card details never touch our servers.
What's included
Secure checkout by Stripe. Card details never touch our servers.
What's included
Secure checkout by Stripe. Card details never touch our servers.
Charges appear as “KAVEL AI” on your card statement. You can cancel any time from Settings → Billing; cancellation takes effect at the end of the period you already paid for.
Powered by
Use one balance across every supported image and video model — ten of them today, listed below — and check the credit cost before each request.
Video models
Image models
Model credit guide
You're only charged for successful generations. The exact estimate in the generator varies by model, length, resolution, audio, and number of images.
| Type | Model | Credit cost |
|---|---|---|
| Video | Seedance 2.0 | 5s 720p image-to-video ≈ 188 credits; text-to-video ≈ 308 credits. Scales with resolution and length. |
| Video | Seedance 2 Fast | Faster and lower cost. 5s 720p text-to-video ≈ 248 credits. |
| Video | Seedance 2 Mini | The cheapest Seedance tier. 5s 720p ≈ 154 credits, 5s 480p ≈ 72 credits. |
| Video | Seedance 1.5 Pro | Audio doubles the rate. 5s 720p ≈ 27 credits silent, ≈ 53 with audio; 1080p ≈ 57 and ≈ 113. |
| Video | Veo 3.1 | Billed per video, not per second (Lite tier). About 45 credits at 720p and 53 at 1080p. |
| Video | Kling 3.0 | Audio raises the rate. 5s 720p ≈ 105 credits silent, ≈ 150 with audio; 1080p ≈ 135 and ≈ 203. |
| Video | MiniMax H3 | Fixed 2K, no resolution ladder. Priced per second — a 5s clip ≈ 158 credits. |
| Image | Nano Banana 2 | Generate or edit from text and images. About 8 credits per 1K image, 12 at 2K, 18 at 4K. |
| Image | Nano Banana Pro | Consistent run times across generations. About 12 credits per 1K or 2K image, 21 at 4K. |
| Image | Nano Banana 2 Lite | Faster, simpler variant. A flat 8 credits per image at every resolution. Start here to test. |
| Image | GPT Image 2 | The lowest-cost image model here. About 3 credits per 1K image, 6 at 2K, 12 at 4K. |
| Image | Seedream 5.0 Lite | Flat pricing — about 8 credits per image whether you generate or edit. |
Common questions about Google's Gemini Flash image tier.
Yes. Gemini Flash Image is the name inside Google's product line; Nano Banana 2 is the name it became known by. Same model, same capabilities, same credit cost.
Pro is the tier to use when text inside the image has to stay legible or when fine detail has to survive enlargement. Flash is the faster, cheaper middle that handles composites and general generation.
Lite takes a single source photo and costs the same at every resolution. Flash takes up to 14 reference photos and lets you choose 1K, 2K or 4K, with the cost rising by tier.
Up to 14 reference photos in one run, accepted as JPEG, PNG or WebP at up to 30MB each.
8 credits at 1K, 12 at 2K and 18 at 4K. The estimate shows before you generate, and a failed generation is refunded automatically unless it broke the content policy.
If you are not sure, start on Lite — it is the free tier's engine, so you can judge real output before spending. Move to Flash the first time you need more than one source photo.
Open the generator, pass text or up to 14 photos, pick a resolution, and generate.
8 credits at 1K, 12 at 2K, 18 at 4K — so the tier you pick is a real decision.
