The generative AI competition is now venturing into the three-dimensional space. Anthropic, World Labs, OpenAI, and ByteDance have recently showcased their related accomplishments, giving rise to three primary technological pathways:
Firstly, the large language model route, exemplified by OpenAI's GPT-6 Astra and Anthropic's Claude Fable 5.1, harnesses the general intelligence of large language models to fuel 3D content creation. These models are integrated into professional software to cater to the needs of creators.
Secondly, ByteDance has developed a real-time spatial video generation model based on Seedance. This model is dedicated to crafting real-time virtual environments that dynamically respond to user actions and sounds, making it ideal for widespread consumption scenarios like live streaming, interactive short dramas, and gaming.
Thirdly, World Labs' Atlas native world model route takes a fundamentally different approach by reconstructing spatial understanding from scratch, with a strong emphasis on spatial consistency. This route is well-suited for applications in film and television production as well as game prototype development.
These three routes vary in their technological underpinnings, capabilities, and industrial orientations, and all currently have areas for enhancement. Looking ahead, they are poised to excel in distinct segments and scenarios, engaging in both competition and integration.