Did Codex Reset
GitHub

Command Code scores three models on 3D SubwayBench

Command Code

Command Code tested Mimo-V2.6-Flash, Grok-4.7, and DeepSeek-V4.1-Flash on 3D SubwayBench using the same /design prompt.

DeepSeek-V4.1-Flash scored 9/10 at $0.024 in one shot, which Command Code described as strong 3D UI/UX and gameplay. Mimo-V2.6-Flash scored 8/10 at $0.018 and produced a playable 2D result in about four iterations. Grok-4.7 scored 3/10 at $0.35; Command Code said its UI and gameplay were slower and needed more iteration.