# Nano Banana 2 > Create and edit with Nano Banana 2 on MixVio. Readable in-image text, subject consistency, owned-reference edits up to 4K—credits shown first, private outputs. - [Canonical Nano Banana 2 page](https://mixvio.ai/models/nano-banana-2) - [Official model source](https://ai.google.dev/gemini-api/docs/image-generation) ## Route summary - **Provider:** Google - **Media:** image - **Inputs:** Text · Image - **Output:** 1K–4K PNG - **Credits:** 5–7 credits per output - **Availability:** Available ## Capabilities - **World knowledge:** Keep explainers, diagrams, and place-specific scenes concrete—so isometric layouts and labeled knowledge boards stay specific instead of generic stock art. - **Text rendering:** Keep magazine covers, recipes, and packaging labels readable—including multilingual and localized copy—so campaign layouts ship without repairing broken letters. - **Precise edits:** Start from an owned product or portrait reference and change only what you ask—background, lighting, or campaign chrome—while identity and label facts hold. - **Subject consistency:** Keep multiple characters and props recognizable in one coherent scene—so cast sheets, product kits, and story worlds stay consistent across a Flash-speed generate-and-edit loop. ## How to use 1. **Start from text or references:** Open Nano Banana 2 with a text brief, or add up to 14 owned images (JPEG, PNG, or WebP, ≤30 MB each) for reference-led create or the image-edit route. Prefill from a Key Feature or Example if you want—nothing runs until you generate. 2. **Spell out text, subjects, and settings:** Write exact headlines, labels, multilingual or localized copy, cast or product identities, materials, and any surgical changes in plain language. Point at specific uploads with @ImageN when needed. Set ratio—including ultra-wide 8:1—and 1K–4K resolution for the delivery you need. 3. **Generate, review, then edit:** Confirm the credit estimate and generate. Check readable text, identity, and layout, then download—or continue on image-edit to change only what you ask while the rest of the frame holds. ## Pricing Generation cost varies by the selected input and output controls. 5–7 credits per output. - [MixVio pricing](https://mixvio.ai/pricing) ## FAQ ### What is Nano Banana 2? Nano Banana 2 is Google DeepMind’s Gemini 3.1 Flash Image model for image generation and editing. On MixVio you can start from a text prompt or owned references, then generate or continue with a natural-language edit—credits shown before you run. ### What is Nano Banana 2 best for? Use it for fast balanced generate-and-edit loops: social and campaign stills, knowledge-grounded explainers, readable in-image text, product restages, and multi-subject scenes where Flash-level iteration speed matters as much as fidelity. ### Can Nano Banana 2 render readable text and localize in-image copy? Yes—text rendering and localization are core strengths. It handles magazine covers, packaging, recipes, and multilingual signage. Still review fine print, spelling, brand names, and dense body text before you publish. ### How many reference images can I use? Up to 14 owned JPEG, PNG, or WebP images per run on MixVio (≤30 MB each). MixVio does not publish a separate pixel-size gate for these stills. Use them for subject lock, product restages, or multi-reference scenes—every file must be yours or cleared for use. ### Can I edit an existing image? Yes. Upload up to fourteen owned images, then describe the change in plain language—for example restage a product, update on-image copy, or change wardrobe. MixVio routes to the Nano Banana 2 edit path. Mask painting is not exposed yet, and unrequested details can still shift, so review before publishing. ### How do I keep characters or products consistent? Attach clear owned references and name them in the prompt with @ImageN tags when you need the model to pull a specific subject, outfit, or product. Keep the same references across generate and edit loops, and re-check identity after each run. ### Which resolutions and aspect ratios does MixVio support? Text-to-image and editing support 1K, 2K, and 4K at 16:9, 9:16, or 1:1—the fixed ratio subset accepted by both active Providers. Auto is not exposed. Desktop starts at 16:9 and mobile starts at 9:16; use 1K to explore, then raise resolution for finals. ### Does Nano Banana 2 use live Google Search on MixVio? No. Live Google Search grounding is not enabled on MixVio’s Nano Banana 2 routes. Rely on your prompt, owned references, and the model’s trained knowledge—then fact-check any dates, prices, or brand claims before publishing. ### How do I get better results? Write exact headlines and labels, attach the references that must stay locked, set ratio and 1K–4K resolution up front, and prefer short edit loops that change only what you ask. Review readable text and identity before you download. ### How many credits does Nano Banana 2 use? From 5 credits per output; resolution and output count (up to four on text-to-image) update the total—5 at 1K, 5 at 2K, and 7 at 4K. MixVio shows the exact estimate before submission. Failed runs cost no credits. ### Is Nano Banana 2 free? Yes on the Free plan when you have credits. Free images are watermarked and capped at 1K. Paid plans unlock higher resolutions without the MixVio watermark. Failed runs cost no credits. ### How does Nano Banana 2 compare with Nano Banana Pro, GPT Image 2.5, or Qwen Image 3? Choose Nano Banana 2 for faster balanced generate-and-edit loops with strong general knowledge. Prefer Nano Banana Pro for detailed product photos and readable type, GPT Image 2.5 for precise keep-and-change edits, or Qwen Image 3 for dense layouts and multilingual text at 1K or 2K. Use the comparison table on this page for live Best for, grades, resolution, and credits. ### Can I use Nano Banana 2 outputs commercially? Paid-plan outputs may be used commercially, subject to the Terms of Service and current Google terms. You must have rights to every prompt, reference, script, face, and brand. Free-plan outputs are for evaluation and are not for commercial publishing. MixVio does not guarantee exclusivity or trademark clearance. ### How are Nano Banana 2 images marked as AI-generated? Nano Banana 2 images include SynthID watermarking from Google. Treat that mark as part of the delivery pipeline and follow current Google / Provider terms when you publish. Free-plan images also carry a visible MixVio watermark; that mark is separate from Google SynthID. ### Are my prompts and outputs private on MixVio? Yes. Prompts are routed server-side, and final media is archived privately to your workspace—not exposed in public URLs or client logs. Keep private prompts and asset IDs out of shared links. ### How does Nano Banana 2 compare with Nano Banana Pro and 2 Lite? Nano Banana 2 is the balanced generate-and-edit choice in the family, with selectable 1K, 2K, or 4K output from 5–7 credits per image. Nano Banana Pro targets detailed product photos and readable type at 7–9 credits. Nano Banana 2 Lite is the low-latency draft route fixed at 1K for 3 credits. Open the Nano Banana Pro page for that pass, or choose Lite in Create.