ChatGPT and Codex limits to be shared starting December; Instant Chat excluded

OpenAI will begin sharing usage limits between ChatGPT and Codex across Plus, Pro, and Business plans starting in December, with the company confirming the change through its official help documentation. Usage will be measured in tokens—the information units AI models consume when processing inputs and generating responses—rather than message counts, meaning consumption varies by model choice and task complexity even for identical request volumes. Instant chat is explicitly excluded from the shared usage pool and will not draw down the combined allowance. While December was given as the rollout month, OpenAI did not specify an exact date or year. The change matters because it consolidates how subscribers draw on capacity across OpenAI's coding and conversational products, potentially shifting effective value per plan tier as heavy Codex users gain flexibility. Watch for the precise implementation date and whether OpenAI adjusts pricing or token allocation rates across tiers once the shared model takes effect.