What's New

Video progress no longer resets during generation. Long prompts and pasted text stay within the editor and result cards.

Improved gallery loading and removed large gaps between pages. Page navigation now shows loading feedback.

If a sign-in attempt fails, you can try again without getting stuck in a loading state.

Seedance 2.5 now checks reference-video requirements before charging credits, so unsupported inputs can be corrected first.

Upload one person photo and 1–8 clothing reference photos to create one outfit. Add your own clothing details to guide the result.

Separate person and clothing upload areas make setup clearer. After generation, drag the comparison slider to check the outfit against the original photo.

Explore 79 new Nano Banana prompt templates, bringing the library to 894. Each new template includes its creator and source, and you can use its prompt to start creating.

Choose GPT Image 2.5 Flare or Sunburst for text-to-image creation and reference-image editing.

Set image size and quality separately, including XHigh and Max quality options. Check the displayed credit cost before generating.

Create 1–15 second videos from text or reference images with MiniMax H3. Choose landscape, portrait, or square format in the video creator.

Create with text, opening and closing frames, or a combination of image, video, and audio references. Choose your duration and resolution in the video creator.

Use the image and video creators, template gallery, and main site pages in Japanese. Switch languages from the site language menu.

The image creator and Midjourney page now offer V8.1 for text-to-image creation.

Create from text, opening and closing frames, or multimodal image, video, and audio references in one focused workspace.

The redesigned model picker and settings controls make it easier to compare Veo 3.1 and Seedance 2.0 options, capabilities, and credit costs.

A new mobile mode selector, settings drawer, responsive model picker, and steadier previews keep the full video workflow comfortable on smaller screens.

The image generator has been completely upgraded with a smoother, more intuitive creation experience.

NanaBanana now supports both light and dark modes, so you can choose the theme that feels most comfortable while creating.

Veo 3.1 and Veo 3.1 Fast are available again for text, reference-image, and first/last-frame video generation.

Create with all nine Standard, Fast, and Mini models using text, frames, or multimodal image, video, and audio references.

Generated videos now appear in a gallery grid right below the creator. Run multiple generations at once and browse all results in one place.

Click any result to open a cleaner, full-size player with easy download — works great on mobile too.

Start from text, guide the result with reference images, or set the opening and closing frames for more control.

The preview area is clearer while your video is generating, and finished videos are easier to review and download as MP4.

Veo 3.1 now fits more naturally into NanaBanana's main video creator, with clearer setup and model choices.

Supported Veo 3.1 modes now make it easier to create from text, reference images, or opening and closing frames, with Standard and Fast choices where available.

Video generation now shows a smoother progress bar and a reminder that you can come back later to check the result in My Creations.

Completed video generations now appear more reliably in My Creations, making finished videos easier to find and download.

Create stunning artwork with Midjourney directly in Nana Banana. 4 unique variations per generation.

Fixed an issue where downloads could occasionally fail.

Completely redesigned image result panel with cleaner, more minimal styling.

New thumbnail navigation at the top for switching between multiple generated images.

Refined loading animation with subtle progress indicator and cleaner appearance.

Sidebar navigation now shows visual highlighting for the current page.

Improved mobile sidebar with better touch interactions and visual feedback.

Fixed activity banner overlapping header issue.

Friendly notification when credits are running low.

Improved image generation quality and reduced empty response rate.

Image detail popup now uses Drawer component for better iPhone experience.

Auto-retry with backup models when primary generation fails.

Homepage features section now displays as horizontal scrolling carousel.

Image upload limit increased to 40MB and up to 14 images per session.

Full Russian localization added to the platform.

Enhanced purchase analytics with actual payment amounts.

New landing page showcasing Nano Banana Pro 4K image generation model.

Browse and use community-shared prompts with one-click generation.

Curated templates for common image editing scenarios.

Frontend converts images to WebP for faster uploads and better quality.

Fixed critical security vulnerability.

Gallery page to browse AI-generated image examples and templates.

Backend system for managing prompt templates.

Generate high-quality videos with Google's latest Veo 3.1 model.

Restore and enhance old or damaged photos with AI technology.

Create stunning videos with OpenAI's Sora2 text-to-video model.

Change clothes on photos with AI-powered virtual try-on technology.

First public release of NanaBanana AI image editing platform.

Core AI-powered image editing with character consistency technology.

Support for English, Spanish, Chinese, and Dutch.