Practical Guide to Using the "Render Images – Modern Models" Feature in Redraw

This guide walks you through the Render Images feature and its modern models, each offering a different combination of realism, speed, creative control, and cost-effectiveness to transform project images into high-quality renders.
The feature is designed to fit different use cases, from those looking for greater architectural fidelity to those who need to generate variations faster and more economically.

You can use Render to:

• Create renders from project images
• Test different visual styles
• Improve the presentation of interior and exterior spaces
• Generate variations more quickly
• Use visual references to guide the result
• Adjust light, climate, scene, and scene elements, depending on the model chosen

1. Access the Platform

Log in to your account at: https://br.arch.redraw.pro/member
If you don't have an account yet, sign up quickly at: https://www.redraw.pro/#pricing

2. Select the Render Images Feature

In the main menu, choose the Render Images feature.
Inside the feature, you can select from different modern AI models based on the type of result you want to achieve.
The main models currently available are:

• Render V4 Lumi
• GPT Image 2
• Nano Banana Pro
• Nano Banana 2

Each model has its own characteristics in terms of performance, visual quality, and creative control.

Render V4 Lumi

The Render V4 Lumi is Redraw's own model, developed to raise the level of realism, spatial reading, and fidelity to the original project.

It brings significant advances in the interpretation of:

• Lines
• Materials
• Depth
• Lighting
• Architectural context

As a result, it tends to better preserve the project's structure, delivering a more realistic image that is coherent with the submitted environment.

Main benefits of Render V4 Lumi:

• More accurate reading of materials, lighting, and depth
• Greater sense of realism, scale, and architectural context
• Higher resolution compared to previous versions
• Redraw's own model at a reduced cost

Additional V4 Features

How does Environment Selection work?

The V4 Lumi lets you choose between two types of environment:

• Interior
• Exterior

This selection defines which scene options become available during generation.

For interior environments, you can use the Light & Climate settings to adjust the scene's lighting.

For exterior environments, in addition to Light & Climate options, you can activate specific settings to better compose the project's surroundings, including:

• Add neighboring houses
• Choose a Scene to define the surroundings
• Adjust Clouds
• Change the Sky color

Scenes can help contextualize the building better, such as street with neighbors, lakeside, rural pasture, lot, eco-condominium, and other options available on the platform.

How do I use Light & Climate?

The Light & Climate section lets you configure the image's atmosphere before generating. You can start from a ready-made scene or set up your configuration manually.

Using a ready-made scene

In the Scenes tab, select one of the available options, such as:

• Midday
• Night
• Sunset

When you choose a scene, Redraw automatically applies a pre-defined lighting and climate configuration.

Adjusting manually

You can also configure each parameter individually:

• Time of day — use the slider to define the angle of light incidence in the scene
• Clouds — control cloud coverage in the sky (available for exterior environments)
• Sky color — click the field to open the color picker and set a specific tone for the sky (also available for exterior environments)

How do I save and reuse presets?

After setting up Light & Climate the way you want, you can save that combination as a preset to reuse in other projects.

How to save a preset

1. Set up the Time of day, Clouds, and Sky color as desired
2. Click + Save preset
3. In the window that opens, give your preset a name (e.g., PROJECT 1)
4. Click Save preset

How to use a saved preset

Go to the Saved tab within Light & Climate. Your presets will be listed, just click one to apply the saved configuration to the current generation.

How do I add objects to the scene?

The V4 Lumi lets you insert objects into the image before generating, directly through the Scene Elements section.

How to add an object

1. In the Scene Elements section, click Add objects
2. The Library will open on the left side of the screen, with items organized by All and Saved
3. Click the object you want to use — it will appear over the image
4. Reposition the object by dragging it across the screen

Object editing options

With an object selected on the canvas, an action bar appears at the bottom with the following options:

• Remove Background — removes the object's background for a cleaner composition
• Flip — mirrors the object horizontally
• Position — adjusts the position in the layer
• Layer — controls the object's overlap in relation to other elements

After positioning and adjusting the objects, click Generate image to include them in the render.

Reference Image

The reference image helps guide the visual result of the generation.

It can be used to indicate:

• Desired style
• Composition
• Lighting
• Materials
• Textures
• Visual identity
• Scene atmosphere

This feature is useful when you want to bring the result closer to a specific reference.

GPT Image 2.5 – Speed and Precision for Generations and Edits

GPT Image 2.5 is the new generation of OpenAI's model available in Redraw, found in both Render and the Text to Image tool. It combines faster generation, sharper details, and greater precision when modifying an existing image, better preserving the reference image and the original composition.
In Render, the model is already prepared to interpret the rendering context. That means no prompt is required: simply upload your image, choose the mode and quality, and generate. The prompt and advanced options remain available whenever you want to guide specific changes or refine the result.

Generation modes

• Fast: prioritizes speed. Recommended for tests, variations, prompt validation, and first versions.
• Precision: prioritizes greater control and fidelity. Recommended for detailed edits, refinements, and images that need to preserve lighting, textures, and composition.

A good strategy is to start with Fast mode to explore ideas and switch to Precision during the refinement stages.

Quality levels

• Economy: quick tests and idea exploration with lower coin consumption.
• Standard: a balance between quality and cost, great for intermediate versions.
• High: more detail and definition.
• Very High: a more refined finish, with attention to textures and element preservation.
• Maximum: prioritizes maximum fidelity, recommended for final work.

All quality levels are available in both modes. The combination you choose determines the coin cost, shown in the interface before generation — Precision mode has a higher coin cost, and the most economical combination is Fast + Economy.

Additional features

Like the other modern models, GPT Image 2.5 is compatible with:

• Reference image: upload up to 2 images to guide the style, composition, lighting, materials, and atmosphere of the scene.
• Manual scene composition (free prompt): describe the desired result in your own words.
• Advanced options: adjustments to mood, atmosphere, environment, and scene elements.

How to generate with GPT Image 2.5 in Render

1. Open Render and upload the image you want to work on (or continue a session).
2. Under Generation model, select GPT Image 2.5.
3. Choose between Fast and Precision modes.
4. Select the desired quality.
5. If you'd like, enable the prompt, reference image, or advanced options to guide the result.
6. Check the coin cost shown in the interface and click Generate image.

Key benefits of GPT Image 2.5:

• Faster generation
• Sharper details, with more natural lighting and textures
• More precise edits, changing parts of the image without losing the other elements of the scene
• Better preservation of the reference image and original composition
• Five quality levels that can be combined with either mode

Nano Banana Pro – Superior Quality with Creative Control

The Nano Banana Pro is a model aimed at those who want higher visual quality and more creative control in render generation.
It has good capacity for scene interpretation, material reading, and fidelity to the architectural context.

Nano Banana Pro features:

• Generation in 2K or 4K
• Free prompt (manual composition)
• Reference image with up to 2 images
• Manual scene composition

Reference Image
You can also send up to two reference images to guide:

• Objects
• Textures
• Lighting
• Composition
• General scene style

Main benefits of Nano Banana Pro:

• Superior quality for architectural renders
• More creative control
• More guided and educational experience
• Ability to work with free prompts and visual references

Nano Banana 2 – Speed and Economy

Indicated for those who want to test more variations, validate ideas, or generate images at a lower coin cost.
Like the other models, you can submit the base project image and generate the render with or without the available advanced options.

Quality options in Nano Banana 2:

• 1K
• 2K
• 4K

Main benefits of Nano Banana 2:

• Faster render generation
• Lower coin cost
• Ideal for testing more variations in the same project
• Good option for initial studies and idea exploration

Available Settings and Features

Depending on the selected model, the feature may include resources such as:

• Image quality
• Advanced options (mood, atmosphere, environment and element adjustments)
• Manual scene composition (free prompt)
• Reference image
• Light & Climate presets

These controls let you adapt the generation to your goal, whether a more technical, commercial, or conceptual image.

Frequently Asked Questions

Do I need to fill in the prompt field to use Render V4 Lumi?
No. The prompt is optional. Redraw's features already have internal settings to guide the generation. Use the prompt only if you want to add specific information about the environment, style, material, or detail you want.
Are the library objects from Redraw, or can I add my own?
The library displays objects you have added to the platform from files on your device. You can also save favorite objects to the Saved tab for easier access in future generations.
Can I use Light & Climate for interior environments?
Yes. The time of day control can be used for both interior and exterior environments. The clouds and sky color options are exclusive to exterior environments.
Are my presets saved permanently?
Yes. Saved presets remain available in the Saved tab and can be reused in future generations.
Which model uses the fewest coins?
Coin consumption may vary depending on the model and quality chosen. In general, faster and more economical models are recommended for testing and variations, while higher-quality models may have a higher cost. Before generating, check the information displayed in the platform interface.
Can I use a reference image with all models?
Reference image availability may vary depending on the chosen model. When the feature is available, it will appear in the feature's interface.
Is GPT Image 2 still available?
Yes. GPT Image 2 remains on the platform, now among the legacy models. Learn more in the Render Images – Legacy Models article.
What is the difference between Fast and Precision modes?
Fast prioritizes speed and is recommended for tests and variations. Precision prioritizes control and fidelity for detailed work, with a higher coin cost.
Do I need to fill in a prompt to use GPT Image 2.5 in Render?
No. You can upload your image, choose the mode and quality, and generate right away. The prompt, reference image, and advanced options are optional.

Conclusion

The Render Images feature in Redraw brings together different AI models to meet varied needs in the render creation process.
With it, you can choose between greater architectural fidelity, more creative control, higher speed, coin savings, or superior quality for presentations — all within Redraw's rendering workflow, allowing you to transform project images into renders in a practical, visual, and personalized way.

Updated on: 17/09/2026