BuildingBench Tests Whether Coding Agents Can Rebuild Real 3D Buildings
Lianhuiq · x · 2026-09-18
EnactraAI introduced BuildingBench, a benchmark testing whether coding agents can turn real-world images into coherent, inspectable and editable 3D buildings for design, games and physical simulation. Early results claim a 80% cost reduction at top quality for one model and a median cost of $2.31 per building for another — though the cited model names ("GPT-6 Astra", "DeepSeek V4.1 Flash") don't match any known releases, so treat with caution. Blog, interactive leaderboard and GitHub are available.
Related event: Enactra Launches BuildingBench for Agent-Driven 3D Building Reconstruction(5 posts)→
More from coding & agent
- Google's open-sourced "Elephant-Goldfish" workflow separates planning from review to curb sloppy AI output — gaganghotra_ · 2026-11-02
- InfoQ: A Systemic Approach to Memory, Knowledge, and Context in Agentic AI Architectures — AdiPolak · 2026-11-02
- Autonomous's Harness Pro uses an ESP32-P4, not a mini PC, to monitor AI coding agents — dee_hw · 2026-11-02
- RL-trained LFM2.5-24B adapter matches accuracy with 20% fewer tool calls — yb2698 · 2026-11-02
- Debugging production AI agents: error surfaces at step 5, root cause at step 3 — Ok-Machine7983 · 2026-11-02
- RAGFlow 1.0 preview: from RAG engine to agent-native database — newpeak · 2026-11-02