← 关卡与可探索世界
工作 · 关卡与可探索世界
生成地形场景与世界内容
截至 2026-10-09，AI 把这项工作做到 L2 部分自动化。3 条更新涉及它，最强的证据：仅来自发布方。
- 级别
- L2部分自动化
- 更新
- 3
- 公司
- 3OpenAI · Google · Meta
- 最强证据
- T3仅来自发布方
L0人工L1辅助L2部分自动化L3有条件自动化L4高度自动化L5完全自动化
是什么把它推到这里
涉及这项工作的全部更新，最新的在前。
- 2026-09-19OpenAIOpenAI showcases 3D builds with GPT-6 Astra
OpenAI shared user-created 3D builds made with GPT-6 Astra.
- 2026-09-03GoogleGemini 3.8 Flash and Antigravity enable complex developer tool building
Google highlighted that Gemini 3.8 Flash and Antigravity make building complex developer tools easier, from procedural game design to multi-platform UI and architectural mapping.
- 2026-08-31MetaImmersive Web SDK builds first-pass 3D/XR scenes from descriptions
Meta described using the Immersive Web SDK, where an agent builds a first pass of a 3D or XR scene in Three.js from a description for refinement in Scene Composer.