Nano Banana 2 (Gemini 3.1 Flash Image)
google/gemini-3.1-flash-imageGoogle’s Nano Banana 2 generates and edits images while accepting image and text input.
131K tokens of context — about 98,304 words, on a scale that ends at 2M.
What it can do
- Reasoning
- Vision
- Takes image
- Takes text
- Gives image
- Gives text
Released 18 Jun 2026 · checked 3 Oct 2026.
About it
Nano Banana 2 is Google’s Gemini 3.1 Flash Image model for generating and editing images. It accepts images and text, can return images and text, and supports reasoning. Its 131,072-token context window handles long inputs, while output is capped at 32,768 tokens. Usage is billed by input and output tokens.
Review
In a line
Nano Banana 2 is Google’s image generation and editing model, also called Gemini 3.1 Flash Image. It accepts image or text input and can return an image or text.
What it's good at
Use it for creating or editing images from text and image prompts. It has vision, so it can work with image input, and it supports reasoning.
The model is part of Google’s Gemini family. See more image generation models and Google models.
Price in plain terms
Input costs $0.5 per million tokens. Output costs $3 per million tokens. Input tokens are what you send; output tokens are what the model returns.
A page of about 500 words costs roughly 0.033 cents to read. A chat turn using about 1,000 input tokens and 500 output tokens costs about 0.2 cents. At that same rate, 1,000 turns cost about $2.
Context and limits
The context window is 131,072 tokens, meaning the model can consider that much material in one request. Its maximum output is 32,768 tokens in one answer.
The listed inputs are images and text, and the listed outputs are images and text. The facts do not specify integrations, a free tier, open-source access, or other limits.
Ways to use it
Use it in software that makes Google’s Gemini 3.1 Flash Image model available for image generation or editing. You can send text, an image, or both, then receive an image or text response.
Its model ID is google/gemini-3.1-flash-image. Compare it with newest models when choosing among newer options.
Verdict
Choose Nano Banana 2 if you need one model that handles image creation, image editing, and text responses, with image understanding and reasoning included. Its context window is 131,072 tokens, and its maximum output is 32,768 tokens.
Skip it if you require a documented free plan, open-source software, named app integrations, or limits beyond those listed here. The facts provide no details on those areas, so buyers should not assume they are available.
Verdict
Nano Banana 2 fits people who want Google’s Gemini image model for creating or changing visuals while still working with text. It can take images as input, return images or text, and supports reasoning. Token pricing is low for ordinary reading and chat volumes, though generated output costs more than input. Consider another model if your choice depends on a free tier, open-source code, specific integrations, or documented provider access, because those details are not given here.
- Generates and edits images from image or text input
- Can return both images and text
- Supports vision and reasoning
- 131,072-token context window
- About $2 for 1,000 typical chat turns
- Output tokens cost more than input tokens
- Maximum output is 32,768 tokens
- No free tier or integrations are specified
- Pick it ifImage generation and editing with text or image prompts; Users comparing Google Gemini models by token cost; Requests that need vision and reasoning
- Skip it ifYou need confirmed app integrations or an open-source model; You require a documented free plan; You need limits beyond the listed context and output caps
Pricing
You pay $0.5 per million input tokens and $3 per million output tokens. A 500-word page costs about 0.033 cents to read; 1,000 typical chat turns cost about $2.
Questions
What is Nano Banana 2?
Nano Banana 2 is Google’s name for Gemini 3.1 Flash Image, an image generation and editing model.
Can it understand images?
Yes. Vision is listed, and image is one of its accepted input types.
What can it produce?
Its listed outputs are image and text.
How much context can it handle?
Its context window is 131,072 tokens, the amount it can consider in one request.
What is the maximum answer length?
The maximum output is 32,768 tokens.
How much does a typical chat turn cost?
A turn using about 1,000 input tokens and 500 output tokens costs about 0.2 cents. 1,000 such turns cost about $2.
What is the model ID?
The model ID is google/gemini-3.1-flash-image.
On these boards
More from Google
- Gemini 3.8 Flash$0.75 in
- Gemini 3.8 Flash (batch)$0.38 in
- Gemini 3.7 Flash$0.75 in
- Gemini 3.7 Flash (batch)$0.38 in
- Gemini 3.5 Flash Lite$0.30 in
- Gemini 3.5 Flash Lite (batch)$0.15 in
- Gemini 3.6 Flash$0.75 in
- Gemini 3.6 Flash (batch)$0.38 in
- Nano Banana 2 Lite (Gemini 3.1 Flash Lite Image)$0.25 in
- Nano Banana Pro (Gemini 3 Pro Image)$2 in
- Gemini 3.5 Flash$1.50 in
- Gemini 3.5 Flash (batch)$0.75 in
- Gemini 3.1 Flash Lite$0.25 in
- Gemini 3.1 Flash Lite (batch)$0.13 in
- Gemini Flash Latest$0.75 in
- Gemini Pro Latest$2 in
- Gemma 4 26B A4B $0.068 in
- Gemma 4 26B A4B (free)Free in