Overview
Nano Banana is a series of AI image generation and editing models launched by Google. It quickly became popular after its initial release in September 2025, and as of May 2026, it has generated over 50 billion images, making it Google's most important product in the consumer-grade AI image field.
Nano Banana Pro (based on Gemini 3 Pro Image, released on 2025-11-20) supports 4K ultra-high definition, advanced text rendering, multi-image fusion (up to 8 images), thinking mode, and web search grounding, capable of transforming recipes, data, real-time weather, and more into well-founded infographics.
Nano Banana 2.1 (gemini-nano-banana-2.1) became generally available on 2026-10-06 as the replacement for Nano Banana 2, with broad gains in image quality, prompt adherence, multi-turn character consistency and text rendering. It outputs at 1K, 2K and 4K with 1K as the default, adds four new panoramic ratios of 1:4, 4:1, 1:8 and 8:1, and fixes tiling artifacts at 2K and 4K. It can blend up to 14 reference images while holding consistency across 4 characters and 10 objects, and supports Google Search grounding plus a configurable thinking level. Nano Banana 2 (gemini-3.1-flash-image) is now deprecated with no shutdown date announced. On 2026-10-08 and 2026-10-09, Artificial Analysis evaluated Nano Banana 2.1 and ranked it 4th on both the AA-Image-T2I v2.0 and AA-Image-Editing v2.0 leaderboards, noting its cost advantage over the previous generation.
Nano Banana 2 Lite: Currently Google's fastest and most cost-effective image generation model, generating images within 4 seconds, costing approximately $0.034 (about 0.23 yuan) per 1000 images, with significantly reduced latency compared to the previous generation. It is suitable for scenarios requiring rapid iterative modifications or batch generation of large volumes of images in a short time. Google recommends that users of the original Nano Banana migrate to this model. It can be used via the Gemini App, Gemini API, and Google AI Studio.
Key Features
- 4K Ultra-High Definition Generation: Nano Banana Pro supports native 4K output, achieving professional-grade image quality; Nano Banana 2.1 also outputs at 1K, 2K and 4K with 1K as the default.
- 2 Lite: 4-Second Image Generation / Ultra-Low Cost: Currently Google's fastest and most cost-effective image generation model, costing only about $0.034 per 1000 images, suitable for high-frequency iteration scenarios.
- Advanced Text Rendering: Clear and accurate text in posters, covers, and infographics, with multilingual support; Nano Banana 2.1 brings broad gains in text rendering.
- Multi-Image Fusion: The Pro version can simultaneously fuse up to 8 reference images to generate coherent and consistent composite works; Nano Banana 2.1 can blend up to 14 reference images while holding consistency across 4 characters and 10 objects.
- Thinking Mode: The Pro version leverages the reasoning capabilities of Gemini 3 Pro to understand complex instructions and adhere to factual constraints; Nano Banana 2.1 supports a configurable thinking level.
- Web Search Grounding: Generates images based on real-time information, such as the latest weather infographics or daily data visualizations; Nano Banana 2.1 supports Google Search grounding.
- Image Editing: Perform language-based editing on existing images, including object replacement, style adjustments, and background modifications.
- Ease of Use and Speed: The Flash series responds quickly, suitable for daily creation and rapid iteration.
- Flexible API Integration: Flexibly call via Gemini API / Vertex AI / third-party proxies (e.g., 0.09 yuan per image).
- Panoramic Aspect Ratios: Nano Banana 2.1 adds four new panoramic ratios of 1:4, 4:1, 1:8 and 8:1, and fixes tiling artifacts at 2K and 4K.
- Google Earth Integration: Since 2026-07-30, the Google Earth web version has offered an image generation feature based on Nano Banana 2, allowing users to combine satellite and 3D imagery with text prompts to reimagine any location worldwide.
Use Cases
- Cover images and supporting graphics for social media content on platforms like WeChat Official Accounts and Xiaohongshu.
- Rapid generation of e-commerce product images and marketing posters.
- Images for PPTs, presentations, and teaching courseware.
- Infographics and data visualizations (leveraging web search grounding capabilities).
- Batch editing of existing images, object replacement, and background changes.
- Quick prototyping for illustrations and concept designs.
- Reimagining locations worldwide by combining satellite and 3D imagery with text prompts in Google Earth.
Pros
- The world's most widely used AI image product (50 billion+ images).
- The Pro version is based on Gemini 3 Pro, with strong understanding of complex Chinese instructions.
- Comprehensive capabilities: 4K high definition + text rendering + multi-image fusion.
- Generous free tier; new users receive tokens across multiple platforms upon registration.
- Affordable API pricing (approximately ¥0.09/image), accessible for small and medium teams.
- Web search grounding enables image generation based on real-time information.
- Nano Banana 2.1 ranked 4th on both the AA-Image-T2I v2.0 and AA-Image-Editing v2.0 leaderboards (2026-10-08, 2026-10-09).
- Nano Banana 2.1 pricing is roughly half of the previous generation for 1K and 2K images.
Pricing
Consumers can try it for free within the Gemini App (with daily usage limits). API usage via Gemini API / Vertex AI is billed per image: Nano Banana Pro costs approximately $0.039 per image, while the Nano Banana Flash series is cheaper. Nano Banana 2.1 is priced at $0.0336, $0.0504 and $0.113 per 1K, 2K and 4K image, roughly half of the previous generation for 1K and 2K. Third-party API proxies offer even lower prices, around ¥0.09 per 4K image. New users enjoy free token quotas in Google AI Studio.
Summary
Nano Banana is the most influential AI image product from 2025 to 2026, thanks to its ease of use, output quality, and Google's continued rapid iteration, most recently with Nano Banana 2.1 and its strong showing on independent leaderboards.
Version History
- Ranked 4th on both Artificial Analysis image leaderboards (2026-10-09): Artificial Analysis finished its independent evaluation of Nano Banana 2.1, placing it 4th on both the AA-Image-T2I v2.0 and AA-Image-Editing v2.0 leaderboards, while noting that it costs half of the previous generation
- Artificial Analysis evaluated Google Nano Banana 2.1, ranking 4th on two leaderboards with a price one-tenth that of the previous generation. (2026-10-08): Artificial Analysis evaluated the image model Nano Banana 2.1 released by Google on October 6, ranking 4th on both the AA-Image-T2I v2.0 and AA-Image-Editing v2.0 leaderboards.
- Nano Banana 2.1 generally available (2026-10-06): Google announced in the Gemini API changelog that Gemini Nano Banana 2.1 (gemini-nano-banana-2.1) is generally available as the replacement for Nano Banana 2, with broad gains in image quality, prompt adherence, multi-turn character consistency and text rendering. Output comes at 1K, 2K and 4K with 1K as the default, four new panoramic ratios of 1:4, 4:1, 1:8 and 8:1 are supported, and tiling artifacts at 2K and 4K are fixed. It can blend up to 14 reference images while holding consistency across 4 characters and 10 objects, and supports Google Search grounding plus a configurable thinking level. Pricing is $0.0336, $0.0504 and $0.113 per 1K, 2K and 4K image, roughly half of the previous generation for 1K and 2K. Nano Banana 2 (gemini-3.1-flash-image) is now deprecated with no shutdown date announced
- Google Earth integrates Nano Banana 2 image generation (2026-07-30): Google Earth web version has launched an image generation feature based on Nano Banana 2, allowing users to combine satellite and 3D imagery with text prompts to reimagine any location worldwide (such as a city's appearance a century ago or a new community sports field). This feature is now available to all users.