Image Generator 2 Chatgpt Image 2.5 & Nano Banana 2
Unik Studio
Image Generator 2.5 FAQ
Common questions about Image Generator 2.5 (v2.5.0), setup, image models, and workflows.
Do I need an API key?
Yes. Generation features require either a Google Gemini / Google AI Studio API key or an OpenAI API key. The add-on purchase does not include provider usage costs.
Does Image Generator 2.5 work offline?
No. This is not an offline renderer. It sends user-started image requests to the selected provider API when you generate, edit, upscale, or check provider access.
Which providers are supported?
Choose Provider, then Model. The v2.5.0 catalog includes four Gemini image models: Nano Banana 2 (Gemini 3.1 Flash Image), Nano Banana Pro (Gemini 3 Pro Image), Nano Banana 2 Lite (Gemini 3.1 Flash Lite Image), and Nano Banana (Gemini 2.5 Flash Image, Legacy). OpenAI offers GPT Image 2, GPT Image 2.5 Sunburst, GPT Image 2.5 Flare, plus legacy GPT Image 1.5, GPT Image 1, GPT Image 1 Mini, and ChatGPT Image. A matching provider API key is required; availability depends on your account. This is an image API catalog, not a list of text-only chat models.
Does the add-on include Gemini or OpenAI credits?
No. Provider billing, quota, account access, model availability, moderation, latency, and returned output are controlled by Gemini or OpenAI. The add-on does not resell or proxy provider access.
Which Blender version is required?
The declared minimum requirement is Blender 4.0 or newer. Package registration has been tested on Blender 5.1; Blender 4.0 has not yet been tested directly.
Where do I find the add-on inside Blender?
Open the 3D Viewport, press N, and select the Image Generator 2.5 tab in the sidebar.
What is Edit Render?
Edit Render uses your current Blender Render Result as the base image for AI editing. It is useful for turning clay renders, product block-outs, scenes, props, and archviz renders into polished visual directions.
Do I need to press F12 before Edit Render?
For the most predictable input, yes: set your camera and press F12 first so Blender has a Render Result to send. The add-on can render when needed, but a deliberate F12 render helps you know exactly what image is being edited.
Can I generate images from text only?
Yes. Switch the mode to Text to Image when you want a prompt-only generation without using a Blender render as the base.
How does masking work?
In Edit Render mode, turn on Mask a region, click Prep & Paint, draw the target area with Blender's Annotate tool, then press IMAGINE - Masked Region. OpenAI uses native mask support where available. Gemini uses add-on-side compositing to keep changes inside the painted region.
What are Image References for?
Image References guide material, palette, tone, lighting mood, surface finish, and overall aesthetic. They are not meant to copy the reference subject or composition. You can add up to five reference images.
What is Iterate on Last Result?
It uses the latest generated image as the next input, so you can keep refining the same image instead of starting again from the original render.
How do I keep the same look across different views?
For Solid Viewport, enable Keep Previous Look. Select a result in History and use Image Ref > Use Selected History to pin it, or choose a loaded Look Ref image. Without a pinned reference, the last generated result is used. The fresh solid capture guides the new viewpoint; the reference guides materials and lighting. Consistency is guided, not guaranteed pixel-identical.
Can I make multiple images from one prompt?
Yes. Use Count to generate multiple sequential variations from one IMAGINE press. Count does not apply to masked edits or AI upscale.
Does 2K or 4K always work?
No. Nano Banana 2 and Nano Banana Pro support 1K, 2K, and 4K requests. Nano Banana 2 Lite and Gemini 2.5 Flash Image use native 1K output. Legacy OpenAI models use 1024x1024, 1536x1024, or 1024x1536 even if you select 2K/4K; 4:5 uses the nearest supported portrait size. GPT Image 2 and 2.5 retain the add-on's output caps. Actual output remains subject to provider and account limits.
What does Upscale (AI) do?
Upscale (AI) asks the selected provider to re-render the selected or latest result with more detail while guiding it to preserve the content and composition. It cannot exceed the model's native output limits, and exact preservation is not guaranteed.
Where are generated images stored?
Generated images appear in the add-on History. From there you can view, save, reuse the prompt, compare with the render, open the output folder, copy the path, clear missing entries, delete entries, or upscale a result.
Why did the provider return no image or an error?
Common causes include missing billing, quota limits, regional restrictions, moderation or safety filtering, unsupported model access, temporary provider issues, network problems, or prompts the provider cannot fulfill. These decisions are controlled by the selected provider.
Does it require Chrome, Edge, or a local browser bridge?
No. Image Generator 2.5 uses provider API keys. It does not require a browser login bridge, Chrome, or Edge for generation.
What is Solid Viewport?
Choose Source > Solid Viewport, frame your subject, and press IMAGINE. The add-on captures a clean solid-shaded view and uses it as the base for an AI image. A free viewport view does not require an active camera or F12 render. Describe the final materials and lighting in your prompt.
What does Legacy mean?
Legacy models remain selectable but have announced retirement dates. The catalog was checked on September 16, 2026: Gemini 2.5 Flash Image is scheduled to stop on October 2, 2026; GPT Image 1 on October 23, 2026; GPT Image 1.5, GPT Image 1 Mini, and chatgpt-image-latest on December 1, 2026. Choose a current model if a legacy endpoint is no longer available. Retired DALL-E, Imagen, and old Gemini preview endpoints are excluded.
Does a ChatGPT subscription replace an OpenAI API key?
No. ChatGPT subscriptions and API billing are separate. Use your own OpenAI API key with image model access and provider billing enabled. The legacy chatgpt-image-latest model does not mean access to every current ChatGPT image feature.
What is the difference between Personal, Team, and Studio?
All three v2.5.0 packages contain the same add-on features. The edition license files retain the previous seat counts: Personal for 1 named user, Team for 5, and Studio for 10. API access and usage fees are separate for every edition.
Can I cancel generation?
Click Cancel or press Esc to stop waiting locally and discard the task's results. Requests already sent to the provider cannot be retracted and may still be processed or billed. Cancellation stops remaining retries and batch requests; late responses cannot overwrite a newer generation.
What support is included?
Support covers installation, Blender UI usage, API key setup inside the add-on, generated image history, masking, upscale, and add-on workflow questions. Provider billing, entitlement, policy, moderation, quota, and service availability are handled by Gemini or OpenAI.
Discover more products like this
Render nano-banana-3 nanobanana openai ai image nanobanana2 chatgpt2.5 image generator nano Generator chatgpt 2.5 chatgpt