Nano Banana 2 Is Here: Gemini 3.1 Flash Image API Access Guide

TL;DR

• Nano Banana 2 is officially launched — it's Gemini 3.1 Flash Image, announced by Google today
• Pro quality at Flash speed — advanced world knowledge, subject consistency (up to 5 characters + 14 objects), precision text rendering, all at Flash-tier speed
• Resolution: 512px to 4K — full range confirmed by Google
• API available in preview — via AI Studio and Gemini API, also on Vertex AI
• EvoLink supports it now — just switch the model name and enjoy better pricing

1. Official Launch: What Google Announced
2. Nano Banana 2 vs Nano Banana Pro vs Nano Banana: Full Comparison
3. What's New in Nano Banana 2
4. Pricing
5. API Integration Guide — Get Started Now
6. Nano Banana 2 Capabilities Showcase
7. Limitations & Workarounds
8. FAQ
9. Join the community challenge — win $1,000 in EvoLink credits
10. References & Update Log

1) Official Launch: What Google Announced

Google officially launched Nano Banana 2 on February 26, 2026. Here's the confirmed rollout:

Now Available ✅

Key Detail: Nano Banana Pro Still Accessible

Google AI Pro and Ultra subscribers can still access Nano Banana Pro for specialized high-fidelity tasks by regenerating images via the three-dot menu in the Gemini app.

2) Nano Banana 2 vs Nano Banana Pro vs Nano Banana: Full Comparison

Why Nano Banana 2 Matters

Nano Banana 2 sits in a sweet spot that didn't exist before:

• For high-volume use cases (e-commerce, social media, ad creatives): You needed Pro quality but couldn't afford Pro latency and cost at scale. Nano Banana 2 solves this.
• For real-time applications (chatbots, interactive tools): Original Nano Banana was fast but quality-limited. Nano Banana 2 brings near-Pro quality at Flash speed.
• For developers: Same Flash-tier pricing, dramatically better output. Your cost per quality-unit drops significantly.

3) What's New in Nano Banana 2

Confirmed by Google's official announcement:

512px to 4K Resolution

Full production-ready resolution range. Create assets for any format — from vertical social posts to wide-screen backdrops — with complete control over aspect ratios and resolutions.

Advanced World Knowledge + Web Search Grounding

The model pulls from Gemini's real-world knowledge base and is powered by real-time information and images from web search. This enables:

• Accurate rendering of specific real-world subjects
• Infographics and diagrams from notes
• Data visualizations
• Factually grounded creative content

Precision Text Rendering & Translation

Generate accurate, legible text in images — and translate/localize text within images to share ideas globally. Game-changer for:

• Marketing mockups and ad creatives
• Greeting cards and invitations
• UI mockups and wireframes
• Multi-language product packaging

Subject Consistency at Scale

• Storyboarding and narrative building
• Brand character consistency across campaigns
• Product catalogs with consistent styling

Enhanced Instruction Following

The model adheres more strictly to complex requests, capturing specific nuances so the image you get is the image you asked for.

Visual Fidelity Upgrade

Vibrant lighting, richer textures, and sharper details — maintaining high-quality aesthetics at Flash speed.

AI Content Provenance

All generated images include SynthID watermarking plus C2PA Content Credentials for transparent AI content identification.

4) Pricing

💡 Via EvoLink: Access Nano Banana 2 with automatic fallback, budget controls, and unified billing across all image models. Check EvoLink pricing.

5) API Integration Guide — Get Started Now

Nano Banana 2 API is live in preview. Here's how to start generating images today.

Why EvoLink for Nano Banana 2?

• Already available: Nano Banana 2 is live on EvoLink
• One-line switch: Just change the model name to gemini-3.1-flash-image-preview — everything else stays the same
• Better pricing: More competitive rates than official pricing
• Automatic fallback: If NB2 has availability issues, EvoLink routes to Nano Banana Pro automatically
• Cost controls: Set budgets and rate limits to manage spend

Step 1 — Get Your API Key

Step 2 — Generate an Image (Python)

import os
import time
import requests

API_KEY = os.environ["EVOLINK_API_KEY"]
BASE_URL = "https://api.evolink.ai/v1"
MODEL = "gemini-3.1-flash-image-preview"

# Step 1: Submit image generation task
response = requests.post(
    f"{BASE_URL}/images/generations",
    headers={
        "Authorization": f"Bearer {API_KEY}",
        "Content-Type": "application/json"
    },
    json={
        "model": MODEL,
        "prompt": "A professional product photo of a smartwatch on a dark reflective surface, studio lighting",
        "quality": "2K",
        "size": "1:1"
    }
)

task = response.json()
task_id = task["id"]
print(f"Task submitted: {task_id}")

# Step 2: Poll for completion
while True:
    status_resp = requests.get(
        f"{BASE_URL}/tasks/{task_id}",
        headers={"Authorization": f"Bearer {API_KEY}"}
    )
    result = status_resp.json()
    if result["status"] == "completed":
        print(f"Image URL: {result['output']['image_url']}")
        break
    elif result["status"] == "failed":
        print(f"Task failed: {result.get('error', 'Unknown error')}")
        break
    time.sleep(3)

Step 3 — Node.js Implementation

const API_KEY = process.env.EVOLINK_API_KEY;
const BASE_URL = 'https://api.evolink.ai/v1';
const MODEL = 'gemini-3.1-flash-image-preview';

// Step 1: Submit task
const taskRes = await fetch(`${BASE_URL}/images/generations`, {
  method: 'POST',
  headers: {
    'Authorization': `Bearer ${API_KEY}`,
    'Content-Type': 'application/json',
  },
  body: JSON.stringify({
    model: MODEL,
    prompt: 'A flat lay infographic explaining the water cycle, clean educational style',
    quality: '2K',
    size: '16:9',
  }),
});

const task = await taskRes.json();
console.log(`Task submitted: ${task.id}`);

// Step 2: Poll for completion
while (true) {
  const res = await fetch(`${BASE_URL}/tasks/${task.id}`, {
    headers: { 'Authorization': `Bearer ${API_KEY}` },
  });
  const result = await res.json();
  if (result.status === 'completed') {
    console.log(`Image URL: ${result.output.image_url}`);
    break;
  }
  if (result.status === 'failed') {
    console.error('Task failed:', result.error);
    break;
  }
  await new Promise(r => setTimeout(r, 3000));
}

Step 4 — cURL

# Submit image generation task
curl -X POST https://api.evolink.ai/v1/images/generations \
  -H "Authorization: Bearer $EVOLINK_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gemini-3.1-flash-image-preview",
    "prompt": "A cyberpunk cityscape at night, neon reflections on wet streets, cinematic",
    "quality": "2K",
    "size": "16:9"
  }'

# Poll task status
curl https://api.evolink.ai/v1/tasks/TASK_ID_HERE \
  -H "Authorization: Bearer $EVOLINK_API_KEY"

Image Editing with Reference Images

Nano Banana 2 supports image-to-image editing by passing reference images:

response = requests.post(
    f"{BASE_URL}/images/generations",
    headers={
        "Authorization": f"Bearer {API_KEY}",
        "Content-Type": "application/json"
    },
    json={
        "model": "gemini-3.1-flash-image-preview",
        "prompt": "Transform this photo into a watercolor painting style",
        "image_urls": [
            "https://example.com/your-photo.jpg"
        ],
        "quality": "2K"
    }
)

Note: Maximum 10 reference images per request (each ≤10MB). Supported formats: JPEG, PNG, WebP. Maximum 5 real person face images.

Async Callback (Optional)

Instead of polling, provide a callback URL to receive results automatically:

response = requests.post(
    f"{BASE_URL}/images/generations",
    headers={
        "Authorization": f"Bearer {API_KEY}",
        "Content-Type": "application/json"
    },
    json={
        "model": "gemini-3.1-flash-image-preview",
        "prompt": "A serene Japanese garden in autumn",
        "quality": "2K",
        "callback_url": "https://your-domain.com/webhooks/image-complete"
    }
)

The callback fires when the task is completed, failed, or cancelled. HTTPS only, max 3 retries on failure.

6) Nano Banana 2 Capabilities Showcase

6.1 Infographics & Data Visualization (NEW)

6.2 Photorealistic Portraits & Characters

Comparison: Nano Banana 2 (left) vs Nano Banana (right) — notice the dramatic improvement in photorealistic quality

6.3 Text-Heavy Content

Nano Banana 2 demonstrating precision text rendering — generating complete, readable text content

6.4 Anime & Stylized Art

Nano Banana 2's anime generation: showcasing perfect action pose and lighting effects

6.5 Creative & Surreal Concepts

7) Limitations & Workarounds

💡 Nano Banana 2's improvements in text rendering and character consistency mean you'll encounter these less frequently than with previous models.

8) FAQ

9) Join the community challenge — win $1,000 in EvoLink credits

10) References & Update Log