I Let GPT-6 Astra and Fable 5.1 Build the Same Game

I put GPT-6 Astra and Fable 5.1 through the same game development challenge to find out how well these AI models can actually build a playable game. I compare their planning, code quality, gameplay, visual results, debugging skills and iteration speed. Most importantly: How much human intervention does each model still need? Which AI is the better game developer—and are we reaching the point where AI can build genuinely good games? Timestamps: 0:00 Intro & Prompt 0:58 Spec files 2:35 Fable 5.1 Game 3:16 GPT-6 Astra 4:27 My 2 Cents