flux-2/max/text-to-image
Create photoreal visuals with multi-reference, color, and typography precision.
Nano Banana 2 Lite is the speed- and cost-focused image model in Google's Gemini image lineup. As a whole, the model both creates images from text and performs conversational edits, with fast multi-turn local touch-ups that keep a scene consistent across rounds. Its design goal is simple: ultra-low latency and a low price so you can generate and refine at volume.
On this RunComfy page you work with the text-to-image side of the model. You send a written prompt, pick an aspect ratio, choose a file format, and receive a fixed 1K (1024x1024) render. The wider editing capability is part of the same underlying model, even though the tool here focuses on generation.
Output format: Resolution: 1K (1024x1024) / fps: n/a / duration: n/a / aspect ratio: auto, 21:9, 16:9, 3:2, 4:3, 5:4, 1:1, 4:5, 3:4, 2:3, 9:16, 4:1, 1:4, 8:1, 1:8 / audio: n/a
| Parameter | Required | Type | Default | Range / Options | Description |
|---|---|---|---|---|---|
| prompt* | Yes (*) | string | — | — | Text description of the image to generate. |
| aspect_ratio | No | string | auto | auto, 21:9, 16:9, 3:2, 4:3, 5:4, 1:1, 4:5, 3:4, 2:3, 9:16, 4:1, 1:4, 8:1, 1:8 | Output aspect ratio; auto lets the model decide. |
| output_format | No | string | png | png, jpeg, webp | File format of the returned image. |
| system_prompt | No | string | (empty) | — | Optional system instruction that steers persona and output style. |
| thinking_level | No | string | off | off, minimal, high | Enable pre-render reasoning; off is fastest, high favors prompt adherence. |
Output is a flat $0.05 per image at 1K (1024x1024). Aspect ratio and file format do not change the price, and generating several images bills per image. Higher thinking_level settings add reasoning work but are covered by the same flat per-image rate.
Create photoreal visuals with multi-reference, color, and typography precision.
Prompt-driven image editing with Nano Banana 2 Edit, with multi-image input plus aspect ratio, resolution, safety tolerance, and output controls.
Edit images with strong prompt control and consistent style using FLUX Kontext Max.
High-fidelity 4-step text-to-image with sharp text rendering
Generate images fast from text with Google Imagen 4 Fast.
Advanced image editing model for detailed, consistent visual creation and precise design workflows.
Nano Banana 2 Lite is the efficiency-focused image model in Google's Gemini image family, built to generate and edit images quickly and cheaply. On this RunComfy page it runs as a text-to-image tool: you describe a scene and it returns a 1K (1024x1024) render. It suits fast ideation, social and marketing visuals, and any workflow that needs many images without a big budget.
Nano Banana 2 Lite is tuned for ultra-low latency, so most prompts return in seconds rather than the longer waits common with heavier premium models. That speed makes it practical for rapid iteration and high-volume drafting. Exact timing varies with load, so treat quick turnaround as the design goal rather than a fixed guarantee.
Output is fixed at 1K (1024x1024 pixels), and you control framing through the aspect_ratio setting. Nano Banana 2 Lite offers 15 options, including 1:1, 16:9, 9:16, 4:3, and 3:4, plus extreme cuts like 4:1, 1:4, 8:1, and 1:8 for banners and strips. You can also leave it on auto to let the model choose a shape that fits the prompt.
As a model, Nano Banana 2 Lite supports both text-to-image generation and conversational image editing with fast multi-turn local touch-ups. The RunComfy tool on this page focuses on the text-to-image side, so it takes a prompt and returns a new image. The broader editing capability is part of the same underlying model family if you plan a wider pipeline.
You can request png, jpeg, or webp. Choose png when you want lossless quality, jpeg for the smallest files, and webp for web delivery that balances size and quality. The format you pick does not change the price or the 1K output resolution.
A text prompt is the only required input, while aspect_ratio, output_format, system_prompt, and thinking_level are optional. Output resolution is fixed at 1K, so plan any upscaling separately if you need larger files. Limits can vary by provider settings, so check the current RunComfy parameter panel before running a large batch.
Yes. You can prototype Nano Banana 2 Lite in the RunComfy model UI and then call the same model via the RunComfy API with identical parameters for production or automation. You don't need to host or scale the model yourself, though you should size your own throughput expectations rather than assume fixed latency.
Generations consume usd or credits, and Nano Banana 2 Lite uses a flat rate of $0.05 per 1K image. Aspect ratio, output format, and thinking_level do not change that flat price, and generating several images is billed per image. See the Generation section of the page for the current rate.
RunComfy is the premier ComfyUI platform, offering ComfyUI online environment and services, along with ComfyUI workflows featuring stunning visuals. RunComfy also provides AI Models, enabling artists to harness the latest AI tools to create incredible art.
