OpenAI lets shared credits cover Codex and other supported features
- OpenAI’s Help Center now says one shared credit balance can fund multiple ChatGPT features, including Codex for Plus and Pro users. - The new docs also put Codex inside spend controls — with add-credits prompts, auto top-up, and a usage dashboard showing balances. - That shifts AI coding from “included perk” toward metered infrastructure teams have to budget, monitor, and govern.
OpenAI is turning ChatGPT credits into a shared wallet instead of a feature-by-feature coupon book. That sounds small, but it changes how people will think about tools like Codex. A coding agent stops feeling like a bundled subscription extra and starts feeling more like cloud compute — something you can burn through, top up, and track. The Help Center updates over the past few days make that much clearer. (help.openai.com) ### What actually changed? The key change is simple — credits are now described as flexible across supported features in the same plan. OpenAI’s consumer help page says those credits can be used across supported features and names Codex for Plus and Pro subscribers. It also says the same balance can show up in the Codex usage flow, so buy(help.openai.com)help.openai.com) ### Why does Codex matter here? Codex is the feature that makes this feel more consequential. Image generation or video already feel like metered compute to a lot of users. A coding agent often does not. But Codex runs real work in the background — generating code, editing files, and(help.openai.com)he Codex help page now frames access by plan and points people back to credits and the rate card for the details. (help.openai.com) ### What do users see when they hit limits? They see pretty standard cloud-software mechanics. OpenAI’s help docs say that if you hit a usage limit in Codex, you get an “Add credits” prompt. Eligible Plus and Pro users can buy more credits from the usage area and can enable auto top-up so the balance (help.openai.com) month.” (help.openai.com) ### Is this only for consumers? No — and the business side is where the model gets even clearer. OpenAI’s flexible-pricing docs for Business, Enterprise, and Edu say shared workspace credits can unlock extra access to advanced features including Codex. Business users get per-seat limits first, then can draw from a shared pool if the works(help.openai.com)es require credits for activity. (help.openai.com) ### What about Enterprise and Edu limits? OpenAI is also tightening the documentation around who gets what by plan. The Enterprise and Edu models-and-limits page now lays out included model access, native tools, and plan behavior in more detail, while related rate-card pages split out c(help.openai.com)— included baseline access on one side, metered overage and advanced usage on the other. (help.openai.com) ### Why package it this way? Because OpenAI is trying to smooth the jump between subscription software and usage-priced AI. If someone already pays for Plus, Pro, Business, or Enterprise, shared credits let OpenAI sell more usage without forcing a plan upgrade every time a user bumps into a limi(help.openai.com)s for Codex, Sora, spreadsheet agents, and whatever comes next. (help.openai.com) ### What’s the bigger implication? The bigger shift is managerial, not technical. Once coding assistance lives inside shared credits, teams have to govern it like any other metered resource. Who gets access? Which models are allowed? How much auto top-up is safe? When does “helpful c(help.openai.com)ard treating AI work as spend that needs controls. (help.openai.com) ### Bottom line OpenAI did not launch a flashy new model here. It did something more structural. It made Codex and other supported tools feel like services drawing from a common budget. That is how cloud platforms behave — and ChatGPT is starting to look a lot more like one. (help.openai.com)
