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

Google has launched Nano Banana 2.1, a new image generation model that costs roughly half as much as its predecessor while delivering…

By Vane October 6, 2026 2 min read
Google’s new image model Nano Banana 2.1 generates better images for less money

Google has launched Nano Banana 2.1, a new image generation model that costs roughly half as much as its predecessor while delivering better results.

Price cuts and model details

The service replaces the February 2026 release of Nano Banana 2. Google states the new version improves quality across the board. Nano Banana Pro likely remains the company’s best image model, though it now uses the Gemini 3.1 engine. The 2.1 update relies on Gemini 3.6 Flash. The next major jump in quality may arrive with the announced Gemini 4 Argon Pro model.

Pricing has dropped significantly. A 1K image now costs 3.36 cents, down from 6.70 cents. A 4K image falls from 15.10 cents to 7.56 cents. The Pro tier remains considerably more expensive at 13.40 cents per 1K image.

Costs for various models and resolutions are as follows:

  • Nano Banana 2.1: 3.36 cents (1K), 5.04 cents (2K), 7.56 cents (4K)
  • Nano Banana 2 / Gemini 3.1 Flash Image: 4.50 cents (0.5K), 6.70 cents (1K), 10.10 cents (2K), 15.10 cents (4K)
  • Nano Banana 2 Lite / Gemini 3.1 Flash Lite Image: 3.36 cents (1K)
  • Nano Banana Pro / Gemini 3 Pro Image: 13.40 cents (1K), 13.40 cents (2K), 24.00 cents (4K)
  • Nano Banana / Gemini 2.5 Flash Image: 3.90 cents (1K)

What the update can do

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, maintaining consistency for up to four characters and ten objects. The model connects to Google Search and offers minimal, medium, and high thinking levels that affect image quality.

Although 2.1 sometimes beats the Pro model 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 results compare the new version against previous iterations:

Overall Preference

  • Nano Banana 2.1 (Thinking): 1050 ± 14
  • Nano Banana 2.1 (No Thinking): 1015 ± 13
  • Gemini 3.1 Flash Image: 990 ± 7
  • Gemini 3 Pro Image: 935 ± 8

Infographic Design

  • Nano Banana 2.1 (Thinking): 1048 ± 17
  • Nano Banana 2.1 (No Thinking): 1001 ± 17
  • Gemini 3.1 Flash Image: 961 ± 12
  • Gemini 3 Pro Image: 912 ± 12

Infographic Accuracy

  • Nano Banana 2.1 (Thinking): 0.521
  • Nano Banana 2.1 (No Thinking): 0.328
  • Gemini 3.1 Flash Image: 0.179
  • Gemini 3 Pro Image: 0.265

General Editing

  • Nano Banana 2.1 (Thinking): 1026 ± 12
  • Nano Banana 2.1 (No Thinking): 980 ± 15
  • Gemini 3.1 Flash Image: 938 ± 11
  • Gemini 3 Pro Image: 939 ± 10

Single-Character Consistency

  • Nano Banana 2.1 (Thinking): 1028 ± 14
  • Nano Banana 2.1 (No Thinking): 1021 ± 14
  • Gemini 3.1 Flash Image: 981 ± 10
  • Gemini 3 Pro Image: 991 ± 9

Multi-Character Consistency

  • Nano Banana 2.1 (Thinking): 1106 ± 14
  • Nano Banana 2.1 (No Thinking): 1068 ± 14
  • Gemini 3.1 Flash Image: 978 ± 10
  • Gemini 3 Pro Image: 1011 ± 10

Mask/Ink-Based Editing

  • Nano Banana 2.1 (Thinking): 1049 ± 15
  • Nano Banana 2.1 (No Thinking): 1042 ± 16
  • Gemini 3.1 Flash Image: 965 ± 12
  • Gemini 3 Pro Image: 927 ± 12

Product Consistency

  • Nano Banana 2.1 (Thinking): 1024 ± 18
  • Nano Banana 2.1 (No Thinking): 981 ± 18
  • Gemini 3.1 Flash Image: 955 ± 22
  • Gemini 3 Pro Image: 965 ± 14

Stylization

  • Nano Banana 2.1 (Thinking): 1062 ± 20
  • Nano Banana 2.1 (No Thinking): 1036 ± 17
  • Gemini 3.1 Flash Image: 991 ± 12
  • Gemini 3 Pro Image: 990 ± 12

Multi-Reference Editing

  • Nano Banana 2.1 (Thinking): 1066 ± 22
  • Nano Banana 2.1 (No Thinking): 1041 ± 20
  • Gemini 3.1 Flash Image: 988 ± 13
  • Gemini 3 Pro Image: 989 ± 12

Availability and rollout

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.

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

Scroll to Top