Copilot is pretty unique in that they were only measuring requests, and that model is just broken for agentic systems.
CC wasn't per-token either? Nor is Codex?
From my simple checks - and from Microsoft's own blog - per token pricing isn't going to be realistic for agentic coding either.
CC wasn't per-token either? Nor is Codex?
From my simple checks - and from Microsoft's own blog - per token pricing isn't going to be realistic for agentic coding either.