Exchange Ranking
Top Exchanges
24h Volume Ranking
Popularity Ranking
Exchange BTC Balance
Proof of Reserves
Decentralized Exchanges
Funding Rate
Funding Heatmap
Liquidation Data
Max Pain
Long/Short Ratio
Whale L/S Ratio
Binance/Okex/Huobi L/S
Bitfinex Margin L/S
ETF Tracker
Solana ETF
XRP ETF
Hong Kong ETF
Bitcoin Treasuries
Crypto Reversal
Ethereum Reserves
HyperLiquid Wallet Analysis
Hyperliquid Whale Watch
Large Transactions
On-chain Movement
Bitcoin ROI
Stablecoin Market Cap
Options Analysis
News
Articles
Economic Calendar
Features
Wallet
Contract Calculator
Security
Collections
Watchlist
FollowingGoogle releases Nano Banana 2.1: AI image generation model ushers in a major upgrade
On Tuesday, Google released Nano Banana 2.1, its latest generation of AI image generation model. The model can convert input text into images or edit existing photos. This update has been launched in Gemini App, Google Search's AI Model, Google Ads, and developer tools such as Google AI Studio, Flow and Stitch.
Market performance and historical background
The original Nano Banana helped Gemini become a phenomenal application. In September 2025, Gemini topped the two major app store rankings with its unique function of turning selfies into collectible hand-made models, ending ChatGPT's nearly three-year leading position. During the same period, the market value of Alphabet (Google's parent company) exceeded the US$3 trillion mark.
The Nano Banana 2 version released in February 2025 is built based on the Gemini 3.1 Flash Image architecture. Its core highlight is the introduction of a real-time retrieval mechanism: before drawing real-world objects, the model will first query Google search results, which significantly improves the accuracy of depicting real events or people.
Nano Banana 2.1 core upgrade
This update focuses on three major improvements:
- Visual design optimization: Overall picture aesthetics and detail expression have been enhanced.
- More accurate mask editing: Users only need to mark a specific area, and only local modifications in this area can be realized, which makes the operation more intuitive and accurate.
- 主体一致性增强：经过多次编辑后，人物或物体的特征仍能保持高度一致，避免形象失真。
性能指标与技术参数
在文本生成图像的偏好测试中，Nano Banana 2.1 取得了 1,050 ELO 分的成绩。相比之下，Nano Banana 2 得分为 990，而 Nano Banana Pro 为 935。在 ELO 评分体系中，分数越高代表表现越优，且该体系无上限限制，纯粹反映相对质量水平。
从技术规格来看，新模型支持同时处理多达 14 张参考图像，能够追踪最多 4 个人物和 10 个物体。输出分辨率最高可达 4K（约等于电视级别清晰度），并支持高达 8:1 的宽屏比例格式。
开发者功能与定价策略
开发者现在可以自定义模型的“思考”时长（从最小到最高不等），并允许模型在生成图像前通过 Google 搜索和 Google 图片搜索进行事实核查。简而言之，模型具备“先查询，后绘图”的能力，以确保内容的准确性。
Through the Google Developer API (an interface that allows other company applications to access the model), standard 1K resolution images cost $0.0336 each, which is only half the price of the Nano Banana 2 ($0.067). The price of a 4K resolution image is US$0.0756, which is also significantly lower than the US$0.151 for the Nano Banana 2. In addition, you can enjoy an additional 50% discount on batch processing tasks (i.e. image collections that are not processed instantly).
At standard 1K prices, the cost of generating a thousand images via the API is now approximately US$33.60, half of the US$67 in the Nano Banana 2 era. This significant cost reduction makes Nano Banana 2.1 not only excellent performance, but also extremely efficient and cost-effective.
