Bijan Bowen reviews GPT-6.1 Terra and testsEvaluation is the systematic assessment of a system's behaviour or results against explicit criteria. it by building a browser operating system and several games. He finds the results fast and generally functional, though often simple, with issues such as inverted controls.
Bijan Bowen demonstrates follow-up prompts that improve the skateboarding and shooter games. A robot-arm exercise stops before completion after the model encounters visual and mapping uncertainty.
Bijan Bowen also tries computer-useComputer use lets an AI agent interact with software through interface controls such as screens, buttons, keyboards and pointers. exercises and a Blender scene. He reports mixed results, including an unsuccessful attempt to refine the scene manually, and offers his assessment of speed, usage and practical value.
Watch on YouTube
