跳到正文
原文
The Decoder· Matthias Bastian·· 3 小時前精選AI 評分76

Google 發佈圖像模型 Nano Banana 2.1,生成效果提升且價格下降

Google's new image model Nano Banana 2.1 generates better images for less money

AI 導讀

Google 發佈圖像生成與編輯模型 Nano Banana 2.1,取代 2026 年 2 月推出的 Nano Banana 2,並稱新版本在多方面均有改進。

推薦理由

文章並列 Nano Banana 2.1 的定價、基準測試與實際生成圖片比較,呈現其成本變化及相對 Nano Banana Pro 的畫質差異。

正文 · 原文

Google has released Nano Banana 2.1, a new model for image generation and editing.

It replaces Nano Banana 2, released in February 2026, and Google says it improves on previous versions "across the board." Nano Banana Pro likely remains Google's best image model, but it uses the most recent Pro model, 3.1, while 2.1 uses Gemini 3.6 Flash. The next major gain in image quality could come with Google's announced Pro model, Gemini 4 Argon.

Google is also cutting prices roughly in half compared with Nano Banana 2. A 1K image costs 3.36 cents, down from 6.70 cents, while a 4K image drops from 15.10 to 7.56 cents. Nano Banana Pro costs considerably more at 13.40 cents per 1K image.

Model 0.5K 1K 2K 4K
Nano Banana 2.1 - 3.36 5.04 7.56
Nano Banana 2 / Gemini 3.1 Flash Image 4.50 6.70 10.10 15.10
Nano Banana 2 Lite / Gemini 3.1 Flash Lite Image - 3.36 - -
Nano Banana Pro / Gemini 3 Pro Image - 13.40 13.40 24.00
Nano Banana / Gemini 2.5 Flash Image - 3.90 - -

According to Google, version 2.1 improves visual quality, text rendering, character consistency across conversation turns, wide panoramas, and infographics. It can process up to 14 reference images at once, keeping up to four characters and ten objects consistent. The model also connects to Google Search and offers minimal, medium, and high thinking levels that affect image quality.

Although 2.1 sometimes beats Pro by a wide margin in benchmarks, Nano Banana 2 also matched it in those tests. In practice, Nano Banana Pro often still produces noticeably better images. Google describes 2.1 as "the more efficient counterpart" to Pro.

Benchmark Nano Banana 2.1 (Thinking) Nano Banana 2.1 (No Thinking) Gemini 3.1 Flash Image Gemini 3 Pro Image
Text-to-Image
Overall Preference 1050 ± 14 1015 ± 13 990 ± 7 935 ± 8
Infographic Design 1048 ± 17 1001 ± 17 961 ± 12 912 ± 12
Infographic Accuracy 0.521 0.328 0.179 0.265
Editing
General Editing 1026 ± 12 980 ± 15 938 ± 11 939 ± 10
Single-Character Consistency 1028 ± 14 1021 ± 14 981 ± 10 991 ± 9
Multi-Character Consistency 1106 ± 14 1068 ± 14 978 ± 10 1011 ± 10
Mask/Ink-Based Editing 1049 ± 15 1042 ± 16 965 ± 12 927 ± 12
Product Consistency 1024 ± 18 981 ± 18 955 ± 22 965 ± 14
Stylization 1062 ± 20 1036 ± 17 991 ± 12 990 ± 12
Multi-Reference Editing 1066 ± 22 1041 ± 20 988 ± 13 989 ± 12

The model runs at Flash-tier speed, supports 1K, 2K, and 4K resolutions, and is already rolling out across Google's products. These include the Gemini app, AI Mode in Google Search, Google AI Studio, Flow by Google, Stitch by Google, Google Ads, and the Gemini Enterprise Platform.

This is the better of two images generated with Nano Banana 2.1 in Google Flow, and it follows our benchmark prompt fairly closely. "Widescreen 16:9, a hyper-realistic DSLR photo. A monkey holding a pink banana is sitting on a tiger in the foreground. In the background, a HORSE is RIDING AN ASTRONAUT. The astronaut is underneath, like a living 'spacesuit horse saddle,' and the HORSE is clearly on top, in control, as the rider. Make it 100% unambiguous: the HORSE is the rider and the ASTRONAUT is being ridden—NOT the other way around. High-resolution, sharp focus, realistic lighting"
Side by side, Nano Banana Pro's image still looks more realistic and natural, particularly in its colors and proportions. Nano Banana 2.1 struggles with scale, making the horse look more like a pony.

More details are available in the model card. Google will shut down the previous model, "gemini-3.1-flash-image," on October 29, 2026.

AI News Without the Hype – Curated by Humans

Subscribe to THE DECODER for ad-free reading, a weekly AI newsletter, our exclusive "AI Radar" frontier report six times a year, full archive access, and access to our comment section.

來源:The Decoder · the-decoder.com