Nano Banana 2.1

Quick example

curl https://reapi.ai/api/v1/images/generations \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "nano-banana-2.1",
    "prompt": "A clean infographic poster titled \"HOW COFFEE IS MADE\" with four labeled steps",
    "aspect_ratio": "3:4",
    "resolution": "2k"
  }'

import requests

resp = requests.post(
    "https://reapi.ai/api/v1/images/generations",
    headers={
        "Authorization": "Bearer YOUR_API_KEY",
        "Content-Type": "application/json",
    },
    json={
        "model": "nano-banana-2.1",
        "prompt": 'A clean infographic poster titled "HOW COFFEE IS MADE" with four labeled steps',
        "aspect_ratio": "3:4",
        "resolution": "2k",
    },
    timeout=30,
)
print(resp.json())

const r = await fetch("https://reapi.ai/api/v1/images/generations", {
  method: "POST",
  headers: {
    Authorization: "Bearer YOUR_API_KEY",
    "Content-Type": "application/json",
  },
  body: JSON.stringify({
    model: "nano-banana-2.1",
    prompt: 'A clean infographic poster titled "HOW COFFEE IS MADE" with four labeled steps',
    aspect_ratio: "3:4",
    resolution: "2k",
  }),
});
console.log(await r.json());

package main

import (
    "bytes"
    "encoding/json"
    "fmt"
    "io"
    "net/http"
)

func main() {
    body, _ := json.Marshal(map[string]any{
        "model":        "nano-banana-2.1",
        "prompt":       `A clean infographic poster titled "HOW COFFEE IS MADE" with four labeled steps`,
        "aspect_ratio": "3:4",
        "resolution":   "2k",
    })
    req, _ := http.NewRequest("POST",
        "https://reapi.ai/api/v1/images/generations", bytes.NewReader(body))
    req.Header.Set("Authorization", "Bearer YOUR_API_KEY")
    req.Header.Set("Content-Type", "application/json")

    resp, _ := http.DefaultClient.Do(req)
    defer resp.Body.Close()
    out, _ := io.ReadAll(resp.Body)
    fmt.Println(string(out))
}

Authentication

Every call needs a Bearer token. Generate keys at reapi.ai/settings/apikeys.

Authorization: Bearer YOUR_API_KEY

Endpoint

POST /api/v1/images/generations
GET  /api/v1/tasks/{id}

Submission is async. The POST returns immediately with a task id; the task endpoint returns the same envelope until completion. Polling does not consume credits. A typical image is ready in about 35–45 seconds — poll every 3–5 seconds and allow a few minutes before giving up.

Request body

Mode is implicit: no image_urls / video_urls is text-to-image; references turn it into editing, reference-guided generation, multi-image fusion or video-to-image.

Checked before anything is submitted or charged:

• prompt is present
• aspect_ratio is auto or one of the 14 ratios above (9:21 and pixel sizes are rejected)
• resolution is 1k, 2k or 4k — there is no 0.5k on this model
• prompt is at most 262,144 characters
• image_urls has at most 14 entries and video_urls at most 10, all public HTTP(S) URLs
• thinking_level is minimal, medium or high
• No n, seed, temperature or top_p — one image per request

google_search / google_image_search — boolean, default false

Grounding lets the model look things up before drawing — today's weather for a forecast card, a real landmark, the current look of a product. Turn on google_search for web results, google_image_search for reference images from Google image search, or both. Google does not use real-world images of people from web search for grounding.

When Google returns search suggestions for a grounded request, they come back as an HTML snippet in output.search_suggestions. Google requires you to display these search suggestions when you use image search grounding, and its Gemini API terms set the full display rules for grounded results. Render the snippet as-is (for example in a sandboxed iframe) next to the image. Not every grounded result carries one.

thinking_level — string, default medium

How much the model reasons about the prompt before generating. minimal is the fastest; high spends more reasoning on dense layouts, infographics and long text. Leaving it out is the same as medium. The level does not change the price.

Troubleshooting

See the errors catalog for every code.

Response envelope

Submit and poll share the same shape — only status and output fill in over time.

{
  "id": "task_018f5a3a1b6e7d9f8c2b4d6e8f0a2c4e",
  "model": "nano-banana-2.1",
  "status": "completed",
  "created_at": 1735000000,
  "output": {
    "image_urls": ["https://cdn.reapi.ai/media/tasks/.../0.jpg"],
    "search_suggestions": "<style>…</style><div class=\"container\">…</div>"
  },
  "error": null
}

Poll GET /api/v1/tasks/{id} (see the Tasks reference) until status === "completed". Images are JPEG. search_suggestions appears only on grounded results that include it. Generated URLs expire — mirror the files to your own storage.

Pricing

Nano Banana 2.1 bills one flat price per image:

credits = ceil(per_image_price × 1000)     1 credit = $0.001

The live per-image rate is on the model page. Failed tasks are always refunded in full.

1k, 2k and 4k cost the same, and so do reference images and videos, grounding and every thinking level. That makes 2k or 4k the sensible choice whenever you need the detail.

Tips

• Spell out the exact text you want in the image, in quotes (titled "HOW COFFEE IS MADE") — text rendering and infographic layout are the areas Google improved most in this version.
• For multi-image fusion, say which reference is which in the prompt ("the mug from the first image on the beach from the second").
• Wide panoramas (4:1, 8:1) and tall strips (1:4, 1:8) at 2k and 4k: Google fixed the tiling artifacts that Nano Banana 2 showed at these ratios.
• Use thinking_level: "high" for infographics, menus and dense layouts; minimal for quick drafts.

Related

• Nano Banana 2 — the previous version
• Nano Banana Pro
• Nano Banana 2 Lite
• Tasks reference
• Errors catalog