

Nano Banana Pro
#21 in Reasoning ModelsUnknown · nano pro · since 20. November 2025 · 2× · last seen Sep 20, 2026
Nano Banana Pro (official name: Gemini 3 Pro Image) is Google DeepMind's advanced image generation and editing model built on Gemini 3 Pro. It was released on November 20, 2025, and features improved text rendering in images, native 4K resolution, the ability to blend up to 14 reference images, and integration with Google Search for fact-grounded graphics. The model is available via the Gemini app, Google AI Studio, Vertex AI, and the Gemini API (model ID gemini-3-pro-image-preview), and achieves top rankings in image benchmarks such as the LMArena image-editing leaderboard. All generated images carry a SynthID watermark.
Features
| Key Benchmark (%) | #1 on LMArena Image Editing leaderboard (1249 Elo); #3 Text-to-Image (1273 Elo) |
| Context Window (Tokens) | 1,000,000 tokens (underlying Gemini 3 Pro base model) |
| License | Proprietary (Google, paid API access) |
| Multimodality | Text and image input/output; can blend up to 14 reference images and connect to Google Search |
| Platform | Gemini app, Google AI Studio, Vertex AI, Gemini API, Google Ads, Workspace (Slides, Vids, NotebookLM), Antigravity |
| Price per 1M Tokens | Image output: $0.134 per image (1K/2K), $0.24 per image (4K) via Gemini API |
| Release Date | November 20, 2025 |