Per-image prices for Nano Banana Pro and Nano Banana 2 from each provider's own public page, read on 2026-09-29. VELIN is included and is operated by the publisher of this page; where others are cheaper, we say so.
Disclosure: VELIN, the service in the highlighted rows, is operated by the publisher of this page. Every other price below comes from that provider's own public page, read on 2026-09-29. Where a price was unclear, we left it out. Prices change often; use the links to check the current figure.
Sorted from lowest to highest 1K price. "n/a" means the provider's page did not show a clear price for that size.
| Provider | 1K | 2K | 4K | Notes |
|---|---|---|---|---|
| VELIN (operated by the publisher of this page) | $0.037 | $0.037 | $0.037 | ¥0.25 flat at every size. Async only, about 45 to 100 s. |
| Google Gemini API, Batch / Flex | $0.067 | $0.067 | $0.12 | Half price, not real-time. source |
| Kunavo | $0.067 | $0.067 | $0.12 | Listed as "$0.067+" per image. |
| kie.ai | $0.09 | $0.09 | $0.12 | Credit price at $0.005 per credit (18 / 24 credits); a ~10% top-up bonus lowers it. |
| Hypereal | $0.10 | n/a | n/a | One price listed for "Nano Banana Pro"; no size split shown. |
| PiAPI | $0.105 | n/a | n/a | One price listed for nano-banana-pro; no size split shown. |
| Google Gemini API, Standard | $0.134 | $0.134 | $0.24 | source. Output only; input tokens extra. |
| WaveSpeed | $0.14 | $0.14 | $0.24 | Table on the model page. The same page shows a live price of "$0.126 per run" and says documentation prices may be out of date. |
| Atlas Cloud | from $0.14 | n/a | n/a | Page shows $0.14 per run; per-size prices not shown. |
| Replicate | $0.15 | $0.15 | $0.30 | Per output image, from the page's pricing data. |
| fal.ai | $0.15 | n/a | $0.30 | Page: $0.15 per image, 4K "charged at double the standard rate" (so $0.30). 2K not stated. |
Nano Banana Pro is Google's Gemini 3 Pro Image model. The flat VELIN price is lower than every other figure in this table, at every size. That does not say anything about speed or reliability, and other providers may serve faster or sync responses.
Sorted from lowest to highest 1K price.
| Provider | 1K | 2K | 4K | Notes |
|---|---|---|---|---|
| Google Gemini API, Batch | $0.034 | $0.050 | $0.076 | Half price, not real-time. source |
| VELIN (operated by the publisher of this page) | $0.037 | $0.037 | $0.037 | ¥0.25 flat at every size. Async only, about 45 to 100 s. |
| kie.ai | $0.04 | $0.06 | $0.09 | Credit price (8 / 12 / 18 credits). The page says top-up bonuses bring it to about $0.036 / $0.054 / $0.081. |
| Hypereal | $0.04 | $0.06 | $0.09 | Listed as "$0.040/0.5K/1K, $0.060/2K, $0.090/4K". |
| Kunavo | $0.0469 | $0.0707 | $0.1057 | Shown as 30% off Google's list. |
| APIMaster | $0.0469 | $0.0707 | $0.1057 | Lowest of several routes on the page; other routes list $0.064 to $0.079 at 1K. |
| PiAPI | $0.06 | $0.08 | $0.12 | Pay-as-you-go from a credit balance; monthly plans are optional. |
| Google Gemini API, Standard | $0.067 | $0.101 | $0.151 | source. Output only; input tokens extra. |
| Replicate | $0.067 | $0.101 | $0.151 | Same as Google's list. |
| WaveSpeed | $0.07 | $0.105 | $0.14 | Table on the model page. |
| fal.ai | $0.08 | $0.12 | $0.16 | Page: $0.08 at 1K, 2K at 1.5x and 4K at 2x. The 2K and 4K figures are calculated from those multipliers. |
| Atlas Cloud | $0.08 | $0.12 | $0.16 | Table on the model page. |
Where others are cheaper: at 1K, Google's own Batch price ($0.034) is lower than VELIN's $0.037, and kie.ai says its bonus-adjusted price is about $0.036. kie.ai and Hypereal are $0.003 above VELIN at 1K before any bonus. VELIN's advantage is at 2K and 4K, where its flat price is below every other figure listed. Smaller models are cheaper again: Google lists Nano Banana 2 Lite at $0.0336 per 1K image (Batch $0.0168), and Kunavo lists it at $0.0255. Those are different models and are not in the tables.
VELIN is asynchronous only: you submit a task and poll for the result, which usually takes about 45 to 100 seconds. That suits queues, batch jobs and n8n-style workflows, but not a user waiting on a page load. WaveSpeed's docs describe submit-and-poll with an optional sync mode on some models. APIMaster's page lists both a sync and an async endpoint. Google's Batch and Flex tiers are half price in exchange for not being real-time.
VELIN refunds the reserved credit automatically when a task fails. Kunavo's site says failed requests are not billed. Atlas Cloud's page notes that content-review failures are still partly billed by Google. If you generate many images with edgy prompts, this matters more than a fraction of a cent.
VELIN accepts USDT only for now (cards are coming soon), with no subscription and no automatic free credits. Kunavo lists Stripe checkout with card, Apple Pay, Alipay and WeChat Pay, and a $10 starting top-up. APIMaster lists USDT or Epay. PiAPI has pay-as-you-go credits plus optional monthly plans. If you need a card or an invoice today, that alone may decide it.
VELIN's default is 8 single-image generations per minute per account, up to 14 reference images per request (8 MB each), and generated images are deleted after 10 hours. Higher limits are available on request. PiAPI's pricing page lists concurrent-job limits that rise with the plan. If you need high burst throughput, check this first.
VELIN has its own small REST API and is not OpenAI-compatible. Kunavo and APIMaster say their APIs are OpenAI-compatible, which helps if your tooling already speaks that format. Google's own API is the reference implementation with the widest feature set, including features third parties may not expose.
GET /api/models.Not included: any provider whose pricing page we could not read or whose price was unclear. If you spot an outdated figure, message us on Telegram and we will check the linked page and update it. See also VELIN's own Nano Banana API pricing page.
On the pages read on 2026-09-29, Nano Banana 2 at 1K ranged from $0.034 (Google Batch) to $0.08 (fal.ai, Atlas Cloud), and at 4K from $0.076 (Google Batch) to $0.16. Nano Banana Pro at 1K ranged from $0.067 to $0.15, and at 4K from $0.12 to $0.30. VELIN charges a flat $0.037 (¥0.25) at every size. VELIN is run by the publisher of this page.
Google's own Batch tier at $0.034 per image is the lowest figure in the table, but it is not real-time. Among the others, VELIN ($0.037), kie.ai ($0.04) and Hypereal ($0.04) are close together. kie.ai says top-up bonuses can bring its price to about $0.036. At 1K the differences are a few thousandths of a dollar, so speed, reliability and payment options may matter more than price.
It depends on your sizes. At 1K, per-size providers can be as cheap or slightly cheaper. At 2K and 4K, a flat price is lower than every other figure in the table, because most providers charge more for larger images.
Google bills by output tokens and gives Batch and Flex tiers at half price. Resellers and aggregators set their own per-image prices, often as credits, with bonuses on larger top-ups. Some list several routes at different prices for one model.
The pages describe them as Google's Nano Banana models, but we cannot verify what runs behind each API. Run the same prompt on a small test budget before you commit, and check speed, error rates and output quality yourself.
Policies differ. VELIN reserves credit at submit time and refunds it automatically if the task fails. Kunavo's site says failed requests are never billed. Atlas Cloud's page warns that requests that fail Google's content review are still partially billed by Google. Read each provider's terms before relying on this.
VELIN takes USDT (card payments are coming soon). Kunavo's site lists card, Apple Pay, Link, Alipay and WeChat Pay through Stripe. APIMaster's site lists USDT or Epay. Other providers offer their own options; check the billing page of each.
We tried to make it so: every figure comes from the provider's own public page, each is linked with the date it was read, unclear prices are left out, and rows where others are cheaper are kept in. But VELIN is operated by the publisher of this page, so treat this as one vendor's comparison and check the linked pages yourself.
Want a free test, higher rate limits or volume pricing? Message @ayan8866e on Telegram. Free tests are on request; there are no automatic free credits.
Nano Banana and Gemini are trademarks of Google. GPT Image is a trademark of OpenAI. VELIN is an independent service and is not affiliated with or endorsed by either company. Competitor prices are quoted from the linked public pages on the date shown and may change.