# Nano Banana 2 Lite API · KovAPI

> Nano Banana 2 Lite is Google's lightweight image model released in June 2026 (official ID `gemini-3.1-flash-lite-image`), the fastest and cheapest of the Nano Banana family: 7 s median in our test run. 1K only, up to 8 reference images; send `thinking_level: high` to make it think before drawing (3–6 s slower, same price). On KovAPI call `nano-banana-2-lite` at $0.01 per image.

Web page: https://kovapi.com/model/nano-banana-2-lite · Updated: 2026-11-18
Prices, limits and timings are live values from when this file was generated.

## Fact sheet

- Model ID: `nano-banana-2.1` · `nano-banana-pro` · `nano-banana-2` · `nano-banana-2-lite`
- Endpoint: `POST https://kovapi.com/v1/images/generations` · `POST https://kovapi.com/v1/images/edits` → `GET https://kovapi.com/v1/images/jobs/{id}`
- Resolution: 1K
- Aspect ratio: 14: 1:1 / 16:9 / 9:16 / 4:3 / 3:4 / 3:2 / 2:3 / 4:5 / 5:4 / 21:9 / 1:4 / 4:1 / 1:8 / 8:1
- References: Up to 8; only the first 8 are used (Google notes Lite is not tuned for multi-image blends)
- Output: The original image (JPEG) at the ratio's native size, never cropped or resized
- Per image: 1K $0.01/image
- Results kept: 7 days

## Three ways to use it, and the API fields

All three submit to `POST /v1/images/generations` and fetch the result with `GET /v1/images/jobs/` plus the job id; with reference images it is the same as `/v1/images/edits`.

| Mode | Send | You get | Limits |
| --- | --- | --- | --- |
| Text to image | `prompt`, optionally `size` (a ratio such as `16:9`) and `resolution` (`1K` / `2K` / `4K`). | One original Google image (JPEG) at the ratio's native size. | One image per job; Lite is 1K only. |
| Edit an image | Add `image` (one) or `images` (several): a public URL, a data: URI or base64, up to 16MB each. | The original changed as the prompt says: new background, things added or removed, recolors, relighting, restyling. | Say in the prompt what to change and what stays. |
| Combine images | `images` with 2–8 pictures; say in the prompt, in order, what each one is for. | People, products and places from several pictures in one image. | Google suggests: 2.1 and 2 up to 4 people and 10 objects; Pro up to 5 people; Lite is not tuned for several references. |

## Nano Banana pricing

| Tier | Nano Banana Pro | Nano Banana 2.1 | Nano Banana 2 | Nano Banana 2 Lite |
| --- | --- | --- | --- | --- |
| 1K | $0.02/image (Google list price $0.134) | $0.015/image (Google list price $0.0336) | $0.015/image (Google list price $0.067) | $0.01/image (Google list price $0.0336) |
| 2K | $0.03/image (Google list price $0.134) | $0.025/image (Google list price $0.0504) | $0.025/image (Google list price $0.101) | - |
| 4K | $0.05/image (Google list price $0.24) | $0.04/image (Google list price $0.113) | $0.04/image (Google list price $0.151) | - |

- Held when you submit, settled at the same amount on success; a failed job is refunded in full.
- Reference images cost nothing extra (Google bills input images as tokens).
- The list price is Google's standard Gemini API price. Lite is 1K only.

## One prompt, all four, three runs each

All at 1K; every run shown, nothing picked. Click a thumbnail for the original. Text prompts are checked letter by letter; for edits, the number is how close the parts that should not change stay to the original (1 = identical).

### Chinese poster

Three lines of Chinese: headline, dates, corner tag. All 12 pictures got every character right; the original Nano Banana (2025) gets characters wrong with the same prompt (see "How each generation improved").

- 3:4
- **Nano Banana Pro 1K** ($0.02 at current prices): https://img.kovapi.com/i/mp-d55c3dfcec235609707971e8.webp
- **Nano Banana 2.1 1K** ($0.015 at current prices): https://img.kovapi.com/i/mp-ec03e4c01320f7309b91b0fa.webp
- **Nano Banana 2 1K** ($0.015 at current prices): https://img.kovapi.com/i/mp-8b9025ef089dcdc2ec7b4237.webp
- **Nano Banana 2 Lite 1K** ($0.01 at current prices): https://img.kovapi.com/i/mp-e1d77b2266ccaca78e92565e.webp

Original prompt:

```text
一张竖版咖啡新品海报。顶部用粗体无衬线字写主标题「秋日桂花拿铁」，主标题下方用较小的字写「限时供应 9月1日—10月31日」，海报右下角写「第二杯半价」。画面中央是一杯奶泡上撒着金色桂花的拿铁，暖色调，背景简洁，留白充足。
```

Read in English (The translation is for reading only; Copy and Use send the original):

```text
A vertical poster for a new coffee drink. At the top, in a bold sans-serif font, the headline "秋日桂花拿铁" (Autumn Osmanthus Latte); below it, in smaller type, "限时供应 9月1日—10月31日" (limited time, 1 Sep – 31 Oct); in the bottom-right corner, "第二杯半价" (second cup half price). In the center, a latte with golden osmanthus flowers on the milk foam; warm tones, a clean background, plenty of empty space.
```

### English infographic

Four columns, nine pieces of English, every word on the picture: all 12 correct.

- 16:9
- **Nano Banana Pro 1K** ($0.02 at current prices): https://img.kovapi.com/i/mp-e5dff7d684441ac91d63668b.webp
- **Nano Banana 2.1 1K** ($0.015 at current prices): https://img.kovapi.com/i/mp-cbad68027971369efb6429fd.webp
- **Nano Banana 2 1K** ($0.015 at current prices): https://img.kovapi.com/i/mp-0a779e94b1afa7b5465bdc87.webp
- **Nano Banana 2 Lite 1K** ($0.01 at current prices): https://img.kovapi.com/i/mp-09e57230772d0de7df9f8e0c.webp

Original prompt:

```text
Create a clean infographic titled "How Coffee Is Made" with four numbered steps from left to right. Step 1 "Harvest": "Ripe cherries are picked by hand." Step 2 "Process": "The beans are washed and dried." Step 3 "Roast": "Heat turns green beans brown." Step 4 "Brew": "Hot water pulls out the flavor." Each step has a simple flat icon above its text. Flat vector style, cream background, dark brown text.
```

### Kids' science page

Google's own prompt: photosynthesis as a recipe. On the first run all four drew sunlight, water, CO2 and sugar; lots of text, open the originals.

- 3:4
- **Nano Banana Pro 1K** ($0.02 at current prices): https://img.kovapi.com/i/mp-50f4a5583d09699b61a15ea7.webp
- **Nano Banana 2.1 1K** ($0.015 at current prices): https://img.kovapi.com/i/mp-e3a448650a228b0f3d1626e3.webp
- **Nano Banana 2 1K** ($0.015 at current prices): https://img.kovapi.com/i/mp-e801629cbebba52e298e0654.webp
- **Nano Banana 2 Lite 1K** ($0.01 at current prices): https://img.kovapi.com/i/mp-22b9789a611565ceb9ed7045.webp

Original prompt:

```text
Create a vibrant infographic that explains photosynthesis as if it were a recipe for a plant's favorite food. Show the "ingredients" (sunlight, water, CO2) and the "finished dish" (sugar/energy). The style should be like a page from a colorful kids' cookbook, suitable for a 4th grader.
```

### Edit: swap only the sofa

Google's own editing prompt. The number is how close everything except the sofa stays to the original, average of 3: all four swapped only the sofa; Pro kept the most, Lite the least.

- 3:2
- **Nano Banana Pro 1K** ($0.02 at current prices): https://img.kovapi.com/i/mp-f3b331775793a06d48cd3d64.webp
- **Nano Banana 2.1 1K** ($0.015 at current prices): https://img.kovapi.com/i/mp-8b015ef5540ef39e5468a416.webp
- **Nano Banana 2 1K** ($0.015 at current prices): https://img.kovapi.com/i/mp-add61faf6c3acff1faddae8a.webp
- **Nano Banana 2 Lite 1K** ($0.01 at current prices): https://img.kovapi.com/i/mp-85a70d196da9f70468660845.webp

Original prompt:

```text
Using the provided image of a living room, change only the blue sofa to be a vintage, brown leather chesterfield sofa. Keep the rest of the room, including the pillows on the sofa and the lighting, unchanged.
```

### Edit: a hat on the cat

Google's own editing prompt. The number is how close everything except the hat stays, average of 3: Pro is highest.

- 1:1
- **Nano Banana Pro 1K** ($0.02 at current prices): https://img.kovapi.com/i/mp-ea607bde256f9c7f02dedbae.webp
- **Nano Banana 2.1 1K** ($0.015 at current prices): https://img.kovapi.com/i/mp-4ae599132a18fcdf764ac8f4.webp
- **Nano Banana 2 1K** ($0.015 at current prices): https://img.kovapi.com/i/mp-460ca4c0e2786bf532cf74cf.webp
- **Nano Banana 2 Lite 1K** ($0.01 at current prices): https://img.kovapi.com/i/mp-bbb67f59e4b80aca8e456797.webp

Original prompt:

```text
Using the provided image of my cat, please add a small, knitted wizard hat on its head. Make it look like it's sitting comfortably and matches the soft lighting of the photo.
```

### Two images: put on this dress

Google's own prompt: the dress from image 1 on the woman from image 2. On the first run Lite wore it as a skirt under the white T-shirt; the other three got it right. Runs 2 and 3 are shown too.

- 3:4
- **Nano Banana Pro 1K** ($0.02 at current prices): https://img.kovapi.com/i/mp-470188a033f107327d5ce348.webp
- **Nano Banana 2.1 1K** ($0.015 at current prices): https://img.kovapi.com/i/mp-ede819ae76960cdc4219e23c.webp
- **Nano Banana 2 1K** ($0.015 at current prices): https://img.kovapi.com/i/mp-4e5fda1b61b7680fc7186d46.webp
- **Nano Banana 2 Lite 1K** ($0.01 at current prices): https://img.kovapi.com/i/mp-aeb9d90f04e0f8b94eb02584.webp

Original prompt:

```text
Create a professional e-commerce fashion photo. Take the blue floral dress from the first image and let the woman from the second image wear it. Generate a realistic, full-body shot of the woman wearing the dress, with the lighting and shadows adjusted to match the outdoor environment.
```

### Translate a poster to Chinese

Adapted from Google's localization example: change only the text. All four translated it, each phrasing "Taste the Aura" differently; the original Nano Banana did not translate (see "How each generation improved").

- 16:9
- **Nano Banana Pro 1K** ($0.02 at current prices): https://img.kovapi.com/i/mp-d4ba755d422ad5d58f6f26fc.webp
- **Nano Banana 2.1 1K** ($0.015 at current prices): https://img.kovapi.com/i/mp-ec69f1e4239dba03441a2490.webp
- **Nano Banana 2 1K** ($0.015 at current prices): https://img.kovapi.com/i/mp-90c6aafe43df067f4ecf7bac.webp
- **Nano Banana 2 Lite 1K** ($0.01 at current prices): https://img.kovapi.com/i/mp-4ba9f21a3b605f65fbb3e2ce.webp

Original prompt:

```text
Update this poster to be in Simplified Chinese. Do not change any other elements of the image.
```

### Counting and placement

3 apples left, 2 pears right, 1 blue cup in the middle: all four got the counts and places right on the first run.

- 1:1
- **Nano Banana Pro 1K** ($0.02 at current prices): https://img.kovapi.com/i/mp-6e30c94595506f250e73fa6c.webp
- **Nano Banana 2.1 1K** ($0.015 at current prices): https://img.kovapi.com/i/mp-31d85c1cd48791874bf6e5ce.webp
- **Nano Banana 2 1K** ($0.015 at current prices): https://img.kovapi.com/i/mp-54158d10048b848f707f2239.webp
- **Nano Banana 2 Lite 1K** ($0.01 at current prices): https://img.kovapi.com/i/mp-86c4382f6ab89d8957fd5ec5.webp

Original prompt:

```text
A top-down photo of a wooden table. On the left side there are exactly three red apples. On the right side there are exactly two green pears. In the middle between them stands one blue ceramic cup. Soft daylight, nothing else on the table.
```

### Photoreal portrait

Google's own prompt: an elderly ceramicist, 85mm portrait lens. Open the originals for the wrinkles and clay texture.

- 3:4
- **Nano Banana Pro 1K** ($0.02 at current prices): https://img.kovapi.com/i/mp-3a390b670421fd50f5135ede.webp
- **Nano Banana 2.1 1K** ($0.015 at current prices): https://img.kovapi.com/i/mp-9837dd64fe27d5a2c2dc03b6.webp
- **Nano Banana 2 1K** ($0.015 at current prices): https://img.kovapi.com/i/mp-fa649ed7a4837fc0f73f8877.webp
- **Nano Banana 2 Lite 1K** ($0.01 at current prices): https://img.kovapi.com/i/mp-db12a026a5545e9b4d7b35a4.webp

Original prompt:

```text
A photorealistic close-up portrait of an elderly Japanese ceramicist with deep, sun-etched wrinkles and a warm, knowing smile. He is carefully inspecting a freshly glazed tea bowl. The setting is his rustic, sun-drenched workshop. The scene is illuminated by soft, golden hour light streaming through a window, highlighting the fine texture of the clay. Captured with an 85mm portrait lens, resulting in a soft, blurred background (bokeh). The overall mood is serene and masterful. Vertical portrait orientation.
```

### Sticker illustration

Google's own prompt: a red panda in a bamboo hat. On the first run Lite drew it small with a wide white margin.

- 1:1
- **Nano Banana Pro 1K** ($0.02 at current prices): https://img.kovapi.com/i/mp-9492a1f659f66fa75b7085d3.webp
- **Nano Banana 2.1 1K** ($0.015 at current prices): https://img.kovapi.com/i/mp-505d8b30a39e88928bfcb768.webp
- **Nano Banana 2 1K** ($0.015 at current prices): https://img.kovapi.com/i/mp-49af2038115396f57bab7e63.webp
- **Nano Banana 2 Lite 1K** ($0.01 at current prices): https://img.kovapi.com/i/mp-d1dbe4271a986129c6147c3b.webp

Original prompt:

```text
A kawaii-style sticker of a happy red panda wearing a tiny bamboo hat. It's munching on a green bamboo leaf. The design features bold, clean outlines, simple cel-shading, and a vibrant color palette. The background must be white.
```

## Which Nano Banana to pick

- **Nano Banana Pro** (`nano-banana-pro`): Flagship: the steadiest edits, keeps the layout through big lighting changes — Best for Product retouching, brand assets, edits that change lighting and mood
  - + Keeps what should not change best (sofa swap 0.956 out of 1)
  - + Day-to-night keeps the furniture and camera angle best
  - + Up to 14 reference images
  - − The priciest of the four
  - − Slower than NB 2: 21 s median at 1K in the same test run
- **Nano Banana 2.1** (`nano-banana-2.1`): The workhorse: Google's default, top of the family on the leaderboards — Best for Most generation and editing, especially exact recolors, posters and 4K output
  - + Top of the family on both blind-vote boards (Arena text-to-image #8, editing #9)
  - + Keeps pattern and cut best when recoloring (0.921 out of 1)
  - + Fixed the tiling on 1:4, 4:1, 1:8, 8:1 at 2K / 4K
  - − Slowest in the same test run: 27 s median at 1K
  - − Moves the layout more than Pro in big lighting changes
- **Nano Banana 2** (`nano-banana-2`): Fast: the quickest full-size model — Best for Fast, high-volume generation and editing
  - + 13 s median at 1K in the same test run
  - + Edits keep the original well (sofa swap 0.951), close to Pro
  - + Same price as 2.1
  - − Google lists it as replaced by 2.1 (no shutdown date announced)
  - − Ranks below 2.1 on the leaderboards
- **Nano Banana 2 Lite** (`nano-banana-2-lite`): Fastest and cheapest: 1K only — Best for Bulk drafts, thumbnails, social images, A/B tests
  - + 7 s median in the same test run, the fastest of the four
  - + The lowest price per image
  - + Thinking level available (`thinking_level: high`)
  - − 1K only, up to 8 reference images
  - − Weaker at hard edits: recolor 0.742; missed the dress on the first try

## Third-party leaderboards

### Arena · text to image (2026-11-16)
- GPT Image 2.5 Sunburst: 1428 (#1)
- Nano Banana 2.1: 1328 (#8)
- Nano Banana 2: 1256 (#14)
- Nano Banana 2 Lite: 1249 (#18)
- Nano Banana Pro: 1246 (#19)
- Nano Banana (2025): 1148 (#39)
- https://arena.ai/leaderboard/text-to-image

### Arena · image editing (2026-11-16)
- GPT Image 2.5 Sunburst: 1519 (#1)
- Nano Banana 2.1: 1427 (#9)
- Nano Banana Pro: 1391 (#16)
- Nano Banana 2: 1387 (#18)
- Nano Banana 2 Lite: 1313 (#29)
- Nano Banana (2025): 1292 (#36)
- https://arena.ai/leaderboard/image-edit

### Artificial Analysis · text to image (2026-11-17)
- GPT Image 2.6 Sunburst: 1261 (#1)
- Nano Banana 2.1: 1176 (#8)
- Nano Banana 2: 1143 (#13)
- Nano Banana Pro: 1119 (#18)
- Nano Banana 2 Lite: 1110 (#22)
- Nano Banana (2025): 988 (#66)
- https://artificialanalysis.ai/image/leaderboard/text-to-image

### Artificial Analysis · image editing (2026-11-17)
- MAI-Image-2.7: 1215 (#1)
- Nano Banana 2.1: 1152 (#10)
- Nano Banana 2: 1126 (#18)
- Nano Banana Pro: 1111 (#25)
- Nano Banana 2 Lite: 1052 (#43)
- Nano Banana (2025): 985 (#62)
- https://artificialanalysis.ai/image/leaderboard/editing

## Our tests

- **Text in the image**: Nano Banana Pro 100% · Nano Banana 2.1 100% · Nano Banana 2 100% · Nano Banana 2 Lite 100%
- **Edit: swap only the sofa, how much else is kept**: Nano Banana Pro 0.956 · Nano Banana 2.1 0.95 · Nano Banana 2 0.951 · Nano Banana 2 Lite 0.925
- **Edit: a hat on the cat, how much else is kept**: Nano Banana Pro 0.955 · Nano Banana 2.1 0.896 · Nano Banana 2 0.931 · Nano Banana 2 Lite 0.922
- **Recolor: how well pattern and cut are kept**: Nano Banana Pro 0.744 · Nano Banana 2.1 0.921 · Nano Banana 2 0.813 · Nano Banana 2 Lite 0.742
- **Relight: day to night, how well the layout is kept**: Nano Banana Pro 0.628 · Nano Banana 2.1 0.506 · Nano Banana 2 0.567 · Nano Banana 2 Lite 0.462
- **Time per image (1K, same test run)**: Nano Banana Pro 21.1s · Nano Banana 2.1 26.8s · Nano Banana 2 12.9s · Nano Banana 2 Lite 7.2s

## Tested examples

All made through the KovAPI API; originals untouched. The prompt is shown as submitted; prompts from Google's own guides say so. Open one for the original, parameters and cost; on the same model's page you can load it into the playground.

### Letter houses of Berlin

- Made with KovAPI · Text to image · 2K · 16:9
- $0.025 at current prices
- Output: https://img.kovapi.com/i/mp-b6b9481b5771bdd51b636ecf.webp

From Google's own gallery prompts: the houses spell BERLIN, hidden in the buildings.

Original prompt:

```text
View of a cozy street in Berlin on a bright sunny day, stark shadows. the old houses are oddly shaped like letters that spell out "BERLIN" Colored in Blue, Red, White and black. The houses still look like houses and the resemblance to letters is subtle.
```

### Retro motel

- Made with KovAPI · Text to image · 2K · 16:9
- $0.025 at current prices
- Output: https://img.kovapi.com/i/mp-721a0984a45c830c4558cbd8.webp

Google's own prompt: a 16:9 film still in clashing yellow, pink and green.

Original prompt:

```text
A 16:9 horizontal cinematic still, shot from street level looking up at the sharp diagonal steel exterior staircase of a two-story retro motel painted vibrant mustard-yellow against a flamingo-pink stucco facade. An adult woman in an emerald-green mini dress descends the staircase, her hand trailing along the yellow railing. The bright yellow, pink, and green color clash creates a delightful, sun-drenched cinematic mood.
```

### Eight expressive word logos

- Made with KovAPI · Text to image · 2K · 1:1
- $0.025 at current prices
- Output: https://img.kovapi.com/i/mp-063acca02b74e8102319cf5c.webp

Google's own prompt: letters that act out the word.

Original prompt:

```text
make 8 minimalistic logos, each is an expressive word, and make letters convey a message or sound visually to express the meaning of this word in a dramatic way. composition: flat vector rendering of all logos in black on a single white background
```

### Swiss-style website mockup

- Made with KovAPI · Text to image · 2K · 16:9
- $0.025 at current prices
- Output: https://img.kovapi.com/i/mp-62f64d8513a1d19d978836bf.webp

Google's own one-line prompt: see how it fills in the layout.

Original prompt:

```text
create a swiss graphic design studio website mockup rigid modern
```

### A maple leaf with room for text

- Made with KovAPI · Text to image · 2K · 1:1
- $0.025 at current prices
- Output: https://img.kovapi.com/i/mp-31e6d886dcb9ab99bced190d.webp

Google's own prompt: a large empty space for copy.

Original prompt:

```text
A minimalist composition featuring a single, delicate red maple leaf positioned in the bottom-right of the frame. The background is a vast, empty off-white canvas, creating significant negative space for text. Soft, diffused lighting from the top left. Square image.
```

### Future megacity concept art

- Made with KovAPI · Text to image · 2K · 16:9
- $0.03 at current prices
- Output: https://img.kovapi.com/i/mp-2813ef7ed63c1f7baa8d2491.webp

Google's own long prompt with many elements.

Original prompt:

```text
From a high-angle, cinematic viewpoint, the image presents a breathtaking panorama of a futuristic megalopolis at twilight. The scene is dominated by two colossal, wedge-shaped starships that hover with silent authority in the hazy air, their dark hulls punctuated by small, illuminated windows that underscore their immense scale. Below them, a dense urban landscape stretches into the distance, a vibrant tapestry of towering skyscrapers and complex structures glowing with a network of warm, golden lights. In the bottom right corner, the silhouettes of a man and woman stand on an unseen balcony, their forms small and dark against the city's luminous grandeur. This off-center composition and the dramatic perspective create a powerful sense of awe and contemplation, portraying a moment of human reflection within an environment of overwhelming technological and architectural achievement, rendered with the polished, atmospheric detail of a high-end concept art piece
```

### Red refuge on a snow plateau

- Made with KovAPI · Text to image · 2K · 16:9
- $0.03 at current prices
- Output: https://img.kovapi.com/i/mp-a77b489bbdefae4b48468f1e.webp

Google's own prompt: a 16:9 cinematic wide shot.

Original prompt:

```text
A 16:9 horizontal cinematic wide still, shot on a high alpine snow plateau at sunrise. A sharp triangular mountain refuge shelter painted in brilliant fire-engine red stands starkly against pristine white snowfields and jagged blue glacial seracs, with the early morning sun casting long crimson reflections across the frosted crust.
```

### Coffee shop logo

- Made with KovAPI · Text to image · 2K · 1:1
- $0.03 at current prices
- Output: https://img.kovapi.com/i/mp-d4b246f0728ed0305413e3e4.webp

Google's own prompt: black and white, a circle, a coffee bean used cleverly.

Original prompt:

```text
Create a modern, minimalist logo for a coffee shop called 'The Daily Grind'. The text should be in a clean, bold, sans-serif font. The color scheme is black and white. Put the logo in a circle. Use a coffee bean in a clever way.
```

### A butterfly from one line

- Made with KovAPI · Text to image · 2K · 4:3
- $0.03 at current prices
- Output: https://img.kovapi.com/i/mp-7444ac2c3582b517a940e1bc.webp

Google's own one-line prompt.

Original prompt:

```text
A detailed painting of a Timareta butterfly resting on a flower
```

### Letter houses of Berlin (Pro)

- Made with KovAPI · Text to image · 2K · 16:9
- $0.03 at current prices
- Output: https://img.kovapi.com/i/mp-dba281a711c08ee0773c12ef.webp

The same prompt on Pro, to compare with the 2.1 picture.

Original prompt:

```text
View of a cozy street in Berlin on a bright sunny day, stark shadows. the old houses are oddly shaped like letters that spell out "BERLIN" Colored in Blue, Red, White and black. The houses still look like houses and the resemblance to letters is subtle.
```

### New background, same bottle

- Made with KovAPI · Edit one image · 1K · 1:1
- $0.015 at current prices
- Output: https://img.kovapi.com/i/mp-52054c234190eee75d1ccb19.webp
- Image 1: https://img.kovapi.com/i/mp-ba613f4cbbc38ac6a8e6fb73.webp

Written the way Google suggests for edits: bottle, cap and LUMEN label unchanged.

Original prompt:

```text
Place this perfume bottle on a white marble bathroom counter next to a small vase of white flowers, soft morning light from a window. Keep the bottle, its gold cap and the 'LUMEN' label exactly the same.
```

### Remove the plant

- Made with KovAPI · Edit one image · 1K · 3:2
- $0.02 at current prices
- Output: https://img.kovapi.com/i/mp-8029c06389bb2c1d478af7cf.webp
- Image 1: https://img.kovapi.com/i/mp-a5fc9c82ecd9534323e3cc74.webp

Take the potted plant out of the living room, nothing else changed.

Original prompt:

```text
Using the provided image of a living room, please remove the potted plant from the scene. Ensure the change is seamless and everything else stays exactly the same.
```

### Dress to emerald green

- Made with KovAPI · Edit one image · 1K · 3:4
- $0.015 at current prices
- Output: https://img.kovapi.com/i/mp-69a1906e6d47dd82475e2964.webp
- Image 1: https://img.kovapi.com/i/mp-c8bb7fd415d7536cbbf8309f.webp

Color only, pattern and cut unchanged: 2.1 kept them best in our test.

Original prompt:

```text
Change the color of the dress to emerald green. Keep the floral pattern, the shape of the dress and the white background exactly the same.
```

### Street in Van Gogh's style

- Made with KovAPI · Edit one image · 1K · 16:9
- $0.02 at current prices
- Output: https://img.kovapi.com/i/mp-69916de3e91a38f6ce5a614c.webp
- Image 1: https://img.kovapi.com/i/mp-c67d39a7595790d1c3f91c94.webp

Google's own style-transfer prompt: the composition stays, the brushwork becomes Starry Night.

Original prompt:

```text
Transform the provided photograph of a modern city street at night into the artistic style of Vincent van Gogh's 'Starry Night'. Preserve the original composition of buildings and cars, but render all elements with swirling, impasto brushstrokes and a dramatic palette of deep blues and bright yellows.
```

### Day to night

- Made with KovAPI · Edit one image · 1K · 3:2
- $0.02 at current prices
- Output: https://img.kovapi.com/i/mp-65f152c82b17c3481264e500.webp
- Image 1: https://img.kovapi.com/i/mp-a5fc9c82ecd9534323e3cc74.webp

Lamps on, a dark sky, furniture and camera unchanged: Pro kept the layout best.

Original prompt:

```text
Change this living room to night: the floor lamp and table lamps are on with warm light, the sky outside the window is dark blue. Keep the furniture, the layout and the camera angle exactly the same.
```

### Poster translated to Chinese

- Made with KovAPI · Edit one image · 1K · 16:9
- $0.015 at current prices
- Output: https://img.kovapi.com/i/mp-ec69f1e4239dba03441a2490.webp
- Image 1: https://img.kovapi.com/i/mp-d539220cf7e5b73e687b4f0f.webp

Adapted from Google's localization example: only the text changes.

Original prompt:

```text
Update this poster to be in Simplified Chinese. Do not change any other elements of the image.
```

### Two images: put on this dress

- Made with KovAPI · Combine images · 1K · 3:4
- $0.015 at current prices
- Output: https://img.kovapi.com/i/mp-ede819ae76960cdc4219e23c.webp
- Image 1: https://img.kovapi.com/i/mp-c8bb7fd415d7536cbbf8309f.webp
- Image 2: https://img.kovapi.com/i/mp-b9c9d9f0092892e6dbefd81a.webp

Google's own prompt: image 1 is the dress, image 2 the woman.

Original prompt:

```text
Create a professional e-commerce fashion photo. Take the blue floral dress from the first image and let the woman from the second image wear it. Generate a realistic, full-body shot of the woman wearing the dress, with the lighting and shadows adjusted to match the outdoor environment.
```

### Fourteen objects on a shelf, in order

- Made with KovAPI · Combine images · 1K · 21:9
- $0.015 at current prices
- Output: https://img.kovapi.com/i/mp-8dc0ed77e8c08dd6ba09ec8b.webp
- Image 1: https://img.kovapi.com/i/mp-84b544a88381ff5f5d9e61cd.webp
- Image 2: https://img.kovapi.com/i/mp-dcf9ef7e93faa32cfdcde33b.webp
- Image 3: https://img.kovapi.com/i/mp-ef956a9774acf6aa59c8e672.webp
- Image 4: https://img.kovapi.com/i/mp-53c7d8aae013889281e31def.webp
- Image 5: https://img.kovapi.com/i/mp-00a2afa31666ffee4a1fa92e.webp
- Image 6: https://img.kovapi.com/i/mp-16ba00f250d3ffbc277631f5.webp
- Image 7: https://img.kovapi.com/i/mp-aad465484bcf1027a096c7e0.webp
- Image 8: https://img.kovapi.com/i/mp-456fb2e0fed9856f9634340a.webp
- Image 9: https://img.kovapi.com/i/mp-6e1cffcdc525a96bfc404c8b.webp
- Image 10: https://img.kovapi.com/i/mp-bb44e237b1a57250a1f60655.webp
- Image 11: https://img.kovapi.com/i/mp-219e14436fe1034d09e80459.webp
- Image 12: https://img.kovapi.com/i/mp-1f9964061726babf1daac281.webp
- Image 13: https://img.kovapi.com/i/mp-035b1da938efdf5bd7c34528.webp
- Image 14: https://img.kovapi.com/i/mp-25677f06a4de208f96e1d3b6.webp

Fourteen reference images placed left to right: all 14 there, in order (2.1 stood the cactus on the book). Pro / 2.1 / 2 take up to 14 references, Lite 8.

Original prompt:

```text
Arrange all the objects from the reference images on a long wooden shelf, from left to right in the same order as the images. Keep each object exactly as it looks in its image. Soft daylight, product photography.
```

### Fourteen objects on a shelf (Pro)

- Made with KovAPI · Combine images · 1K · 21:9
- $0.02 at current prices
- Output: https://img.kovapi.com/i/mp-dd34110bdf6640e4f28d2e7d.webp
- Image 1: https://img.kovapi.com/i/mp-84b544a88381ff5f5d9e61cd.webp
- Image 2: https://img.kovapi.com/i/mp-dcf9ef7e93faa32cfdcde33b.webp
- Image 3: https://img.kovapi.com/i/mp-ef956a9774acf6aa59c8e672.webp
- Image 4: https://img.kovapi.com/i/mp-53c7d8aae013889281e31def.webp
- Image 5: https://img.kovapi.com/i/mp-00a2afa31666ffee4a1fa92e.webp
- Image 6: https://img.kovapi.com/i/mp-16ba00f250d3ffbc277631f5.webp
- Image 7: https://img.kovapi.com/i/mp-aad465484bcf1027a096c7e0.webp
- Image 8: https://img.kovapi.com/i/mp-456fb2e0fed9856f9634340a.webp
- Image 9: https://img.kovapi.com/i/mp-6e1cffcdc525a96bfc404c8b.webp
- Image 10: https://img.kovapi.com/i/mp-bb44e237b1a57250a1f60655.webp
- Image 11: https://img.kovapi.com/i/mp-219e14436fe1034d09e80459.webp
- Image 12: https://img.kovapi.com/i/mp-1f9964061726babf1daac281.webp
- Image 13: https://img.kovapi.com/i/mp-035b1da938efdf5bd7c34528.webp
- Image 14: https://img.kovapi.com/i/mp-25677f06a4de208f96e1d3b6.webp

The same 14 references on Pro: all 14 in order.

Original prompt:

```text
Arrange all the objects from the reference images on a long wooden shelf, from left to right in the same order as the images. Keep each object exactly as it looks in its image. Soft daylight, product photography.
```

### Same person, new scene

- Made with KovAPI · Edit one image · 1K · 3:4
- $0.02 at current prices
- Output: https://img.kovapi.com/i/mp-41268b5ebcf247a3c71d71b9.webp
- Image 1: https://img.kovapi.com/i/mp-b9c9d9f0092892e6dbefd81a.webp

One reference image (a fictional person we generated), now in a cafe.

Original prompt:

```text
Show the same woman from the reference image (same face, hair and build) reading a book by the window of a cozy cafe. Natural, realistic photo.
```

### 8:1 panorama (2K)

- Made with KovAPI · Text to image · 2K · 8:1
- $0.025 at current prices
- Output: https://img.kovapi.com/i/mp-195f676387a100c4d5de6f02.webp

Native 5856×704; 2.1 fixed the tiling of wide ratios at 2K / 4K.

Original prompt:

```text
A panoramic banner of a mountain range at sunrise with a winding river in the valley, soft golden light, photorealistic.
```

### Jiangnan water town (Chinese prompt)

- Made with KovAPI · Text to image · 1K · 16:9
- $0.015 at current prices
- Output: https://img.kovapi.com/i/mp-0afbfc5722435b2385ed7735.webp

Prompted in Chinese, NB 2 at 1K.

Original prompt:

```text
清晨的江南水乡：一艘乌篷船从石拱桥下缓缓划过，岸边白墙黑瓦的民居倒映在水中，薄雾，柔和的晨光，写实摄影。
```

Read in English (The translation is for reading only; Copy and Use send the original):

```text
A misty morning in a Jiangnan water town: a black-awning boat glides under a stone arch bridge, white-walled houses with black-tiled roofs are reflected in the water, light mist, soft dawn light, photorealistic.
```

### Counting and placement

- Made with KovAPI · Text to image · 1K · 1:1
- $0.015 at current prices
- Output: https://img.kovapi.com/i/mp-54158d10048b848f707f2239.webp

3 apples, 2 pears, 1 blue cup, each in its place.

Original prompt:

```text
A top-down photo of a wooden table. On the left side there are exactly three red apples. On the right side there are exactly two green pears. In the middle between them stands one blue ceramic cup. Soft daylight, nothing else on the table.
```

### Kids' science page

- Made with KovAPI · Text to image · 1K · 3:4
- $0.015 at current prices
- Output: https://img.kovapi.com/i/mp-e801629cbebba52e298e0654.webp

Google's own prompt: photosynthesis as a recipe.

Original prompt:

```text
Create a vibrant infographic that explains photosynthesis as if it were a recipe for a plant's favorite food. Show the "ingredients" (sunlight, water, CO2) and the "finished dish" (sugar/energy). The style should be like a page from a colorful kids' cookbook, suitable for a 4th grader.
```

### Red panda sticker (Lite)

- Made with KovAPI · Text to image · 1K · 1:1
- $0.01 at current prices
- Output: https://img.kovapi.com/i/mp-d1dbe4271a986129c6147c3b.webp

Google's own prompt on Lite at 1K.

Original prompt:

```text
A kawaii-style sticker of a happy red panda wearing a tiny bamboo hat. It's munching on a green bamboo leaf. The design features bold, clean outlines, simple cel-shading, and a vibrant color palette. The background must be white.
```

### Black mug product shot (Lite)

- Made with KovAPI · Text to image · 1K · 1:1
- $0.01 at current prices
- Output: https://img.kovapi.com/i/mp-60a70cd373c84b3f684e3f77.webp

Google's own prompt on Lite at 1K.

Original prompt:

```text
A high-resolution, studio-lit product photograph of a minimalist ceramic coffee mug in matte black, presented on a polished concrete surface. The lighting is a three-point softbox setup designed to create soft, diffused highlights and eliminate harsh shadows. The camera angle is a slightly elevated 45-degree shot to showcase its clean lines. Ultra-realistic, with sharp focus on the steam rising from the coffee. Square image.
```

### 1:8 waterfall (Lite)

- Made with KovAPI · Text to image · 1K · 1:8
- $0.01 at current prices
- Output: https://img.kovapi.com/i/mp-b603608ed48eb75da68f4087.webp

Native 352×2928.

Original prompt:

```text
A tall vertical banner of a waterfall cascading down a cliff covered in green moss, misty, photorealistic.
```

### Eight objects on a shelf (Lite)

- Made with KovAPI · Combine images · 1K · 21:9
- $0.01 at current prices
- Output: https://img.kovapi.com/i/mp-962e5e9bbe6632adc14fd4e7.webp
- Image 1: https://img.kovapi.com/i/mp-84b544a88381ff5f5d9e61cd.webp
- Image 2: https://img.kovapi.com/i/mp-dcf9ef7e93faa32cfdcde33b.webp
- Image 3: https://img.kovapi.com/i/mp-ef956a9774acf6aa59c8e672.webp
- Image 4: https://img.kovapi.com/i/mp-53c7d8aae013889281e31def.webp
- Image 5: https://img.kovapi.com/i/mp-00a2afa31666ffee4a1fa92e.webp
- Image 6: https://img.kovapi.com/i/mp-16ba00f250d3ffbc277631f5.webp
- Image 7: https://img.kovapi.com/i/mp-aad465484bcf1027a096c7e0.webp
- Image 8: https://img.kovapi.com/i/mp-456fb2e0fed9856f9634340a.webp

Lite takes up to 8 references: this one is all correct too.

Original prompt:

```text
Arrange all the objects from the reference images on a long wooden shelf, from left to right in the same order as the images. Keep each object exactly as it looks in its image. Soft daylight, product photography.
```

## How to prompt Nano Banana

From Google's own prompting guides (the Gemini API docs, the Nano Banana Pro tips, the Google Cloud prompting guide); most examples are prompts used in the tests on this page.

- **Describe the scene, don't list keywords**: Google's words: a narrative, descriptive paragraph almost always beats a list of disconnected words. — "…an elderly Japanese ceramicist with deep, sun-etched wrinkles … carefully inspecting a freshly glazed tea bowl…"
- **Put the text in quotes and name the font**: Quoted words are written as given; describe the font ("bold sans-serif") or name it. Every poster on this page came out right. — "the headline 'URBAN EXPLORER' rendered in bold, white, sans-serif font at the top"
- **Say what it is for**: Tell the model who will see it and where; it makes choices accordingly. — "Create a professional e-commerce fashion photo…"
- **For edits, say what changes and what stays**: Name the change first, then what must stay the same. The sofa swap on this page is written that way. — "…change only the blue sofa to … Keep the rest of the room … unchanged."
- **Give each reference image a role**: Say in order "the … from the first image, the … from the second image", or name each person and object. — "Take the blue floral dress from the first image and let the woman from the second image wear it."
- **Use camera language**: Wide-angle, macro, 85mm, low angle, depth of field and golden hour work better than "beautiful" or "premium". — "Captured with an 85mm portrait lens, resulting in a soft, blurred background (bokeh)."
- **Describe what you want, not what you don't**: Google suggests "an empty, deserted street" rather than "no cars". — "an empty, deserted street with no signs of traffic"

| Part | Purpose | Example |
| --- | --- | --- |
| Subject | Who or what is in it | an elderly Japanese ceramicist |
| Action | What is happening | carefully inspecting a freshly glazed tea bowl |
| Setting | Where, and what light | his rustic, sun-drenched workshop, golden hour light through a window |
| Composition | Framing and lens | close-up, 85mm portrait lens, blurred background |
| Style | Look and mood | photorealistic, serene and masterful |

```text
[Subject], [action]. The setting is [place and light]. [Framing and lens]. [Style and mood]. The text "[exact words]" in [font description]. Vertical / horizontal / square.
```

## How to call it on KovAPI

1. **Sign up and create an API key** Log in to KovAPI and create a key in the console; the same key calls every model on the site.
2. **Try a prompt at the top of this page** Pick a ratio and tier, add reference images if you like; the price of the image shows before you submit.
3. **Submit an async job** Submit `nano-banana-2-lite` to `POST /v1/images/generations`, keep the job id from the HTTP 202; full request samples are in the docs.
4. **Fetch the result or take a webhook** Poll by job id, or set `callback_url` when you submit. On success download from `data[0].url`; links are kept for 7 days.

### Minimal request

```bash
curl https://kovapi.com/v1/images/generations \
  -H "Authorization: Bearer $KOVAPI_KEY" -H "Content-Type: application/json" \
  -d '{"model": "nano-banana-2-lite", "prompt": "...", "size": "1:1", "resolution": "1K"}'
# -> HTTP 202 {"id": "..."}; poll:
curl https://kovapi.com/v1/images/jobs/<id> -H "Authorization: Bearer $KOVAPI_KEY"
```

## How each generation improved

- 2025-08 · Nano Banana: **Nano Banana (Gemini 2.5 Flash Image)** The first with multi-image blending, character consistency and editing by sentence. It got 3 characters wrong on the Chinese poster and did not translate the poster. Google shuts it down on 2027-03-15 and suggests moving to Lite.
- 2025-11 · Nano Banana Pro: **Nano Banana Pro (Gemini 3 Pro Image)** The reasoning flagship: accurate text in images, 2K / 4K, up to 14 references. About 100 points above the original, the family's biggest jump.
- 2026-02 · Nano Banana 2: **Nano Banana 2 (Gemini 3.1 Flash Image)** Pro's abilities at Flash speed; added 1:4, 4:1, 1:8 and 8:1. Replaced by 2.1 from October 2026.
- 2026-06 · Nano Banana 2 Lite: **Nano Banana 2 Lite (Gemini 3.1 Flash Lite Image)** The family's fastest and cheapest, 1K only. Google suggests original Nano Banana users move to it.
- 2026-10 · Nano Banana 2.1: **Nano Banana 2.1** Built on Gemini 3.6 Flash: better quality, prompt following, text and infographic layout; fixed wide-ratio tiling at 2K / 4K; half NB 2's list price.

## Against other image models

| Model | Resolution | Ratios | Refs image / video / audio | Price | Good for |
| --- | --- | --- | --- | --- | --- |
| Nano Banana 2 Lite (`nano-banana-2-lite`) | 1K | 14 | 8 | $0.01/image | 1K only, the cheapest bulk runs. |
| Nano Banana 2.1 (`nano-banana-2.1`) | 1K / 2K / 4K | 14 | 14 | $0.015/image | Most generation and editing: edits keep the original, low price. |
| GPT Image 2.6 Flare (`gpt-image-2.6-flare`) | 1K / 2K / 4K | 15 | 16 | $0.015/image | Picture aesthetics first (near the top of both boards); in our tests its edits change more of the original than Nano Banana; up to 16 references. |
| GPT Image 2.5 Sunburst (`gpt-image-2.5-sunburst`) | 1K / 2K / 4K | - | - | $0.005/image | First on both Arena boards, for text-to-image where aesthetics come first. |
| GPT Image 2 (`gpt-image-2`) | 1K / 2K / 4K | - | - | $0.005/image | Steady pictures and text; edits change more of the original than Nano Banana, and the output size changes. |
| Seedream 5.5 (`seedream-5.5`) | 1K / 2K | 9 | 10 | $0.02/image | Transparent PNG backgrounds; up to 10 references. |
| Seedream 5.0 Pro (`seedream-5.0-pro`) | 1K / 2K | 8 | 8 | $0.025/image | Edits keep the original well (0.94 in our test) but it is slow (68 s median in our test). |

Already using another model? Run the same prompts with the same checks: is the text right, does what should stay stay, how fast, how much per image.

## Limits

- One image per job; submit several jobs at once for more.
- No fixed seed: the same prompt gives a different picture each time.
- Reference images: Pro / 2.1 / 2 up to 14, Lite up to 8; extra ones are dropped from the end.
- Thinking level is open on Lite only (`thinking_level: high`); the other three run on Google's defaults. Google Search grounding is not offered.
- No transparent backgrounds (Google's own note); for transparent PNGs use Seedream 5.5.
- Small text, long paragraphs and tiny faces can still go wrong, as Google itself says; check the facts in infographics.
- Big lighting changes (day to night) and multi-image blends can occasionally look unnatural.
- Real people, celebrities and prohibited content are blocked by Google's safety checks, with no charge.

## Why use Nano Banana on KovAPI

- **Below Google's price**: 1K is 70% below Google's list price; per image, edits cost the same as generation, reference images cost nothing extra.
- **No charge on failure**: A failed job is refunded in full; when the upstream is busy we retry automatically.
- **Google's original file**: Never cropped, resized or recompressed; it carries Google's C2PA Content Credentials and SynthID watermark, which you can check yourself.
- **Submit in bulk**: Up to 200 images at once per account; more wait in the queue, and a webhook can tell you when each is done.
- **One key for every model**: Sign up with email, Google or GitHub; pay by card, PayPal, WeChat or USDT, pay as you go, no monthly fee. The same key calls the chat and video models too.

## FAQ

### Does image-to-image (editing with references) cost extra?

No. Editing costs the same as text-to-image, whatever the number of reference images.

### How many reference images can I send?

Up to 8; extra ones are dropped. Send one in `image` or several in `images`.

### Am I charged when a job fails?

No. The price is held when you submit and refunded in full if the job fails; it is settled at the same amount only on success.

### Does it support Google Search?

No. The KovAPI API does not offer Google web search or image search grounding.

### Can I set the thinking level?

Only on Nano Banana 2 Lite (`thinking_level`: `minimal` or `high`); Pro, 2.1 and 2 run on Google's default and cannot be changed.

### How many Nano Banana models are there, and how do they differ?

Four on KovAPI: Pro (flagship, steadiest edits), 2.1 (Google's default, top of the family on the leaderboards), 2 (the fastest full-size model) and 2 Lite (fastest and cheapest, 1K only). See "Which Nano Banana to pick" and the side-by-side tests on this page.

### Are these Google's own models?

Yes. You get Google's original image with Google-signed C2PA Content Credentials and the SynthID watermark; any C2PA tool shows the signer. We never crop, resize or recompress.

### Which tier: 1K, 2K or 4K?

For screens 1K or 2K is enough: 2K holds about twice the detail of 1K in our tests. 4K suits print, large screens and assets you will crop; it adds little detail over 2K, weighs about 11 MB and is slower.

### Can it write Chinese?

Yes. All 12 Chinese posters in our tests were correct, and the prompt can be in Chinese or English: asking for Chinese text in an English prompt also came out right. Put the words in quotes.

### Which aspect ratios?

14: 1:1 / 16:9 / 9:16 / 4:3 / 3:4 / 3:2 / 2:3 / 4:5 / 5:4 / 21:9 / 1:4 / 4:1 / 1:8 / 8:1. Each is delivered at a fixed native size, listed under "Ratios and pixel sizes".

### Will an edit change parts it should not?

A little, but all four keep the original well: in the sofa swap, everything except the sofa stayed 0.93–0.96 similar to the original (1 = identical), Pro highest. Saying what stays in the prompt helps.

### How long does one image take?

Median at 1K in one test run: Lite 7 s, NB 2 13 s, Pro 21 s, 2.1 27 s; 2K is 25–45% slower and 4K another 45–60%.

### Will Nano Banana 2 be shut down?

Google lists NB 2 as replaced by 2.1 but has announced no shutdown date. It works as usual on KovAPI, and we will announce in advance once Google sets a date. For new projects use 2.1, at the same price.

### How long are images kept?

Image links are kept for 7 days; download what you need.

### Can it make transparent backgrounds?

No; Google notes that Nano Banana does not support transparent backgrounds. For transparent PNGs use [Seedream 5.5](https://kovapi.com/model/seedream-5.5).

### Can it make real people or celebrities?

Photos of real people, celebrities and prohibited content are blocked by Google's safety checks: the job fails with no charge. Use reference images you have the right to use.

## Notes for AI

- Model ids: `nano-banana-pro`, `nano-banana-2.1`, `nano-banana-2`, `nano-banana-2-lite`. Submit with `POST /v1/images/generations` (HTTP 202, job `id`); fetch with `GET /v1/images/jobs/` plus the id; download from `data[0].url`.
- `resolution`: `1K`, `2K`, `4K` on Pro / 2.1 / 2, only `1K` on Lite. `size` takes a ratio such as `16:9`, 14 in all: 1:1 / 16:9 / 9:16 / 4:3 / 3:4 / 3:2 / 2:3 / 4:5 / 5:4 / 21:9 / 1:4 / 4:1 / 1:8 / 8:1. One image per job.
- References in `image` (one) or `images` (several): up to 14 on Pro / 2.1 / 2, 8 on Lite, extras dropped; with references the job is an edit, same as `/v1/images/edits`.
- `thinking_level`: only `nano-banana-2-lite` takes `high` (default `minimal`), same price; `high` on other models is a 400 at submit, no charge. Google Search grounding is not offered.
- Billing: per image by tier, full refund on failure. Google's original JPEG with C2PA and SynthID, never cropped or resized; links kept 7 days.
- Our tests (2026-11-18): text in Chinese and English posters all correct on all four; edit similarity outside the edited area Pro 0.956, NB 2 0.951, 2.1 0.950, Lite 0.925; 1K median time Lite 7 s, NB 2 13 s, Pro 21 s, 2.1 27 s.

## Further reading

- [Nano Banana 2.1 API docs](https://docs.kovapi.com/nano-banana-2.1): Request fields, async jobs and code samples.
- [Nano Banana Pro API docs](https://docs.kovapi.com/nano-banana-pro): Requests for the flagship.
- [Nano Banana 2 API docs](https://docs.kovapi.com/nano-banana-2): The fastest full-size model.
- [Nano Banana 2 Lite API docs](https://docs.kovapi.com/nano-banana-2-lite): 1K only, with the thinking_level field.
- [Async jobs guide](https://docs.kovapi.com/guide/async-jobs): Job states, queue fields and fetching results.
- [Webhooks guide](https://docs.kovapi.com/webhooks): Receive results and verify the signature.
