Comparison of 3D modeling and artificial intelligence tools in landscape design visualizations

Today, landscape architecture is undergoing a significant transformation through the adoption of artificial intelligence (AI) technologies. The literature emphasizes how big data analytics and AI-driven solutions provide in-depth analyses and innovative approaches that strengthen sustainability in landscape design. AI has been embraced as an effective means of overcoming the limitations of current design methods, opening up new possibilities for digitization and automation within landscape architecture workflows. These developments underscore the pivotal role that AI plays in reshaping the profession and driving innovation in design practice. In this study, Antalya Atatürk Cultural Park was chosen as the case study area to compare the effectiveness of 3D modeling tools with AI-based tools in landscape visualization. With its diverse plantings, high visitor traffic, and varied spatial configurations, the site offers an ideal setting for evaluating and contrasting multiple modeling and AI techniques. The research began with a literature review and field survey. Fifty photographs were captured on-site, after which an additional 60 images were generated using Blender, Twinmotion, Fooocus AI, and Adobe Photoshop Beta (2025), with all production times meticulously recorded. A panel of 15 experts reviewed these images and selected 20 for further analysis. Concurrently, an online survey was administered to over 400 participants, who rated each image's similarity to the originals and provided demographic data via a five-point Likert scale. The resulting dataset was processed in Orange software to conduct efficiency, cost, and time analyses, thereby quantifying the performance and preference rates of each software and AI tool. The findings reveal that AI-based tools offer clear advantages in creativity, presentation techniques, time management, and cost efficiency. They also show how demographic factors influence user preferences and identify the model parameters most strongly associated with positive evaluations. These insights can inform the selection of optimal technologies for visual presentations in landscape architecture. Finally, among the techniques tested, the semi-automatic, Photoshop-based, AI-assisted method emerged as the preferred approach particularly for its balance of visual quality, production speed, and cost-effectiveness under current conditions. In contrast, although Blender and Twinmotion produced highly detailed and realistic images, their complexity and longer processing times resulted in lower user preference. Meanwhile, the text-based AI application Fooocus AI provided rapid and economical outputs but has yet to achieve the expected level of visual realism.