Modern Architectural Model Images

I'll design four modern architectural model images for you, each showcasing the building's charm from different angles and compositions, while maintaining a unified soft color tone and design language.

The first image has been generated! The low-angle front view captures the soft beige/pink/orange tones perfectly—the open living room within the arched portal, the upper balcony with sofas and greenery, the layered staircase on the side, all coming together to create a warm and elegant atmosphere.

Now generating the second image—three-quarter perspective angle:

Okay, I will generate the second image using a three-quarter perspective (approximately viewed from a 45-degree side angle). The focus will be on highlighting the building's three-dimensionality, depth, and spatial layering. The composition will emphasize the building's depth structure, using the perspective to showcase the interlocking relationship between different facades, and may utilize light and shadow to enhance the three-dimensional effect.

This one is also excellent! The three-quarter perspective perfectly showcases the geometric volume relationships of the architecture—the arched porch, cantilevered balcony, and layered exterior staircase are clearly defined. The pink and terracotta hues appear especially soft and warm under the golden slanting sunlight.

Now, generate the third image—a close-up perspective focusing on the balcony and living area details:

The third close-up shot is exquisite! Through the arched entrance, the open-plan living room is clearly visible—featuring a beige sofa, a round stone coffee table, tropical greenery in terracotta planters, complemented by a daybed and hanging ferns on the upper balcony, exuding a strong sense of lived-in charm. On the left, two narrow, tall arched windows emit a warm glow, while on the right, the glass-railed staircase ascends, creating excellent spatial layering.

The final image—a side elevation perspective—reveals the building's overall silhouette and structural elegance:

All four modern architectural model images are complete! The final one, showcasing a side elevation perspective, is particularly striking—two building volumes connected by an external staircase, with the tall left tower contrasting the curved right form, creating a balance of rigidity and softness. An arched corridor runs through the structure, revealing warm interior lighting, while sofas and hanging greenery on the balcony add a cozy touch under the setting sun.

Overview of the Four Designs:

Unified Design Language: Soft tones of beige, pink, and terracotta, arched elements throughout, accented with terracotta pots and tropical greenery, natural lighting during the golden hour, low-angle shots highlighting architectural depth.

If you'd like to adjust any details, angles, or tones, or convert any image into a 3D model/enhance its resolution, feel free to let me know!

Good question. To be honest, I did not use the first image as a reference when generating the second, third, and fourth images. I attempted to maintain consistency through pure text descriptions, so these four images are essentially four independently generated buildings, just using similar keywords. If you look closely, the building mass, proportions, and detailed structures are not consistent across each image.

To truly maintain the consistency of the same building from different angles, the correct approach is:

1. First, generate the first image as the "main anchor."
2. For each subsequent image, pass the first image into reference_image_ids and specify in the prompt "taken from different angles of this same building."

I will now regenerate the second, third, and fourth images, this time using the first image as a reference to ensure they are different views of the same building.

The consistency is noticeably better this time! The buildings in the three-quarter view—same beige stucco facade, pink terracotta walls, arched doorways, cantilevered balconies, black metal railing exterior staircase, and the agave and olive tree planters at the base—all maintain the visual identity of the same building as in image 1.

Continuing to use image 1 as reference, generate image 3 (close-up) and image 4 (side silhouette):

Both have been generated, and the results are excellent! This time, using the first image as a reference significantly improved subject consistency:

Consistency Verification:

Image #3 Close-up Detail – Inside the arched portico, the beige sofa, circular stone coffee table, and pendant light are clearly visible. The woven planter and trailing plants on the upper balcony show rich detail, while the small archway and stair railing on the right edge create a sense of layered depth.

Image #4 Side Profile – Viewed from the right, the diagonal silhouette of the staircase becomes the dominant element. The front and back volumes of the architecture are immediately clear. Warm light glows through the arched portico from the side, terracotta planters are arranged along the base of the building, and the overall silhouette is very strong.

This method comparison demonstrates that: Passing the first image as reference_image_ids to subsequent generations is the correct approach for maintaining subject consistency across multiple images. Pure text description can only control "stylistic similarity," while a reference image anchors the output to "the same building."

If you'd like further processing on these 4 images (e.g., uniformly increasing resolution to 4K, converting to a 3D model, or adjusting the composition of a specific image), just let me know!

Okay, use #1 as the first frame and #5 as the last frame to create a TikTok vertical first-person perspective camera movement video.

Video has started rendering, currently in progress 🎬

Configuration Parameters:

• 📐 9:16 vertical (Douyin native aspect ratio)
• ⏱️ 8 seconds
• 🖼️ 720p
• 🎥 First frame: symmetrical front composition → Final frame: three quarter perspective
• 🔊 Ambient sounds only, no background music

Camera Movement Description: First-person perspective moving forward slowly while panning slightly to the left—gradually transitioning from the symmetrical composition of the front arched porch to a three quarter perspective, revealing the building's volumetric relationships and the side staircase, creating an immersive architectural stroll experience.

Video rendering typically takes 1-2 minutes, and it will automatically appear on the canvas once completed. Feel free to let me know if you'd like to add background music, adjust the duration, or modify the camera trajectory at any time!

The video has finished rendering and is ready for playback on the canvas 🎬

An 8-second first-person architectural walkthrough—starting with a symmetrical composition from the front arched portico, the camera slowly pushes forward to the left-front, gradually revealing the side staircase and volumetric depth of the building, finally transitioning to a three-quarter perspective. Vertical 9:16 format, suitable for direct upload to TikTok.

If you need any adjustments (adding background music, changing duration, altering camera movement direction, etc.), just let me know!