A Goal is fundamentally different from a normal prompt because it gives the agent a durable completion condition to pursue across multiple turns rather than a single instruction to execute once.
OpenAI's Codex documentation explains that unlike a normal prompt, which is executed once and waits for the next instruction, a Goal attaches a durable target to a thread so the agent keeps working and self-checking until the objective is satisfied.
transcript
OpenAI (Codex documentation): A normal prompt says: do this next thing. A Goal says: keep working until this outcome is true. In a normal request, Codex works through the immediate instruction, reports a result, and waits. With a Goal, Codex has a durable target attached to the thread.