Issue 2026-11-09 · Industry · 研究 · 应用
Einsia releases Frontier3DBench for coding agents
Einsia released Frontier3DBench, which tests whether coding agents can build editable Blender projects and runnable Unreal Engine worlds across 145 tasks in 9 categories, using 11 model-harness configurations. GPT-6.1-sol leads overall and on objects and scenes, while GPT-6-astra leads interactive worlds; all configurations score lower on reference-based reconstruction than on text-driven generation. Across 165 world submissions, local mechanisms scored 83.08 versus 30.26 for stateful workflows.
Read original ↗