Google's Gemini 4 Argon Autonomously Builds a 3D Pagoda Temple Complex in Blender
Google DeepMind has shown its Gemini 4 Argon model independently constructing a detailed 3D environment inside Google Antigravity. In the demonstration, the model built a multi-building pagoda temple complex in Blender, using multi-agent visual critique loops to refine architecture, foliage, and lighting from minimal starting input. Google presents the work as evidence that an AI agent can plan, execute, and refine a sequence of linked 3D tasks rather than producing isolated assets. The announcement does not claim such scenes can be reliably produced for every use case, and it omits details on time, compute, and iteration counts. Argon is slated for Antigravity access for Google AI Ultra subscribers from October 13, 2026, and is also available via the Gemini API at introductory rates of $2 per million input tokens and $10 per million output tokens.
This is an AI-generated summary. ShortSingh links to the original source for the complete article.

Discussion (0)
Log in to join the discussion and vote.
Log in