Google Launches Nano Banana 2.1 With Lower Prices and Higher Quality
Google introduces Nano Banana 2.1 for image generation and editing, offering quality improvements and roughly halved prices compared to the previous version.
Google has officially introduced Nano Banana 2.1, a brand-new model built for both image generation and editing tasks. Serving as the successor to Nano Banana 2, which originally launched in February 2026, the updated system is described by Google as delivering comprehensive enhancements across the board. While Nano Banana Pro remains the company's leading image model, utilizing the newer Pro 3.1 architecture, the fresh 2.1 release is powered by Gemini 3.6 Flash. Meanwhile, Google has noted that future major leaps in visual quality could arrive with the upcoming Pro model known as Gemini 4 Argon.
Alongside the structural updates, Google has dropped pricing by approximately fifty percent relative to Nano Banana 2. Producing a standard 1K image now costs 3.36 cents, which is down from 6.70 cents. For higher-tier outputs, a 4K image has been reduced from 15.10 to 7.56 cents. In comparison, the higher-end Nano Banana Pro remains significantly more expensive at 13.40 cents per 1K image. The model handles Flash-tier speeds, supports 1K, 2K, and 4K resolutions, and is currently deploying across various Google platforms including the Gemini app, AI Mode in Google Search, Google AI Studio, Flow by Google, Stitch by Google, Google Ads, and the Gemini Enterprise Platform. Google has also scheduled the older "gemini-3.1-flash-image" model to shut down permanently on October 29, 2026.
Performance and capabilities
According to official details from Google, version 2.1 elevates overall visual quality, improves text rendering, maintains better character consistency across conversational turns, handles wide panoramas more effectively, and enhances infographics. The system is capable of processing up to 14 reference images simultaneously, successfully tracking up to four characters and ten objects for visual consistency. Furthermore, the model links directly to Google Search and introduces minimal, medium, and high thinking levels that directly influence the final image output.
In various benchmarks, Nano Banana 2.1 occasionally outperforms the Pro tier by a significant margin, though the previous Nano Banana 2 also achieved parity in similar evaluations. Even so, practical usage reveals that Nano Banana Pro frequently delivers visibly superior and more natural images, particularly regarding colors and proportions. Google characterizes the 2.1 release as the more efficient counterpart to the Pro offering. Additionally, the older generation model, "gemini-3.1-flash-image," will be retired by Google on October 29, 2026.
What it means for developers
For builders integrating visual generation into applications, the release of Nano Banana 2.1 brings noticeable cost savings and stronger multi-reference handling. With prices cut nearly in half and support for up to 14 reference images tracking multiple characters and objects, engineering teams can build complex editing and generation pipelines more affordably. Developers looking to experiment with these capabilities can try top AI models cheaply through one API at https://apixoai.online. The inclusion of variable thinking levels also gives technical teams finer control over the balance between generation speed and output fidelity.
Source: Google's new image model Nano Banana 2.1 generates better images for less money — The Decoder. Written by the Apixo team from that report.
One key for Claude, GPT, GLM, DeepSeek and more. Pay per token with crypto.
Get your API key
