$ cat wiki/models/nano-banana-2-lite.md
Nano Banana 2 Lite (Gemini 3.1 Flash Lite Image)
modelupdated 2026-07-21created 2026-07-14
Spec
| Attribute | Value |
|---|---|
| Developer | Google DeepMind |
| Official name | Gemini 3.1 Flash Lite Image |
| Codename / nickname | Nano Banana 2 Lite |
| Released | 2026-06-30 |
| Announced | 2026-06-30 |
| Context window | unknown |
| Speed | ~4 seconds per image |
| Pricing | ~$0.034 per image |
| Text-to-image Elo | 1,255 (rank #5, Artificial Analysis) |
| Availability | Gemini API, Google AI Studio |
| License | Proprietary (API access) |
Release Date
June 30, 2026 — the fastest and most cost-efficient model in the Nano Banana (Gemini image generation) family. Part of the Gemini 3.1 Flash model family.
The Nano Banana Family (as of July 2026)
| Model | Speed | Quality | Cost |
|---|---|---|---|
| Nano Banana (standard) | Fast | Medium | Low |
| Nano Banana 2 (full) | Medium | High | Medium |
| Nano Banana Pro (high quality) | Slow | Highest | High |
| Nano Banana 2 Lite | Fastest (4s) | Good | Lowest ($0.034) |
Benchmarks
- Text-to-image Elo: 1,255 (rank #5 globally per Artificial Analysis, as of late June 2026)
- Preferred over competing models 72% of the time on complex layouts (multi-column documents, tables with merged cells, text+figures)
Use Cases
- High-throughput API image generation (cheapest Google option per image)
- Rapid prototyping and creative workflows requiring speed over maximum quality
- Mobile and edge deployment scenarios (Flash-class latency)
Compared To
| Model | Speed | Text-to-Image Elo | Price/image |
|---|---|---|---|
| Nano Banana 2 Lite | ~4s | 1,255 (#5) | $0.034 |
| Nano Banana 2 (full) | ~8s | ~1,310 (#3 est.) | ~$0.08 |
| Midjourney v7 | ~15s | ~1,380 (#1 est.) | ~$0.10 |
| Fal.ai FLUX Pro | ~6s | ~1,340 (#2 est.) | ~$0.05 |