OpenAI’s $100 Pro Tier

OpenAI just put a new price tag on heavy artificial intelligence coding: $100 a month buys a ChatGPT plan with 5 times more Codex usage than the $20 Plus tier, and OpenAI says it built the tier for people running longer coding sessions every day. Codex is OpenAI’s coding agent, which means it does more than answer a question in a chat box: it can write code, review code, and work through software tasks across local tools and the web. OpenAI’s help center describes it as a tool to help users “write, review, and ship code faster.” The pricing gap before this was awkward. TechCrunch reported a ladder that jumped from free, to an $8 Go plan, to $20 Plus, and then up to a $200 Pro plan, so the new $100 tier lands exactly in the middle for people who had outgrown Plus but did not want the top bill. OpenAI’s own math explains why the middle tier exists. Its Codex rate card says average usage runs about $100 to $200 per developer per month, with the total moving around based on model choice, number of running instances, automations, and use of fast mode. That makes the new plan less like a luxury add-on and more like a prepaid bucket for a certain kind of worker. If your day is asking an artificial intelligence agent to inspect pull requests, patch bugs, and keep several coding jobs running at once, $20 monthly limits can feel like a commuter rail pass for someone trying to fly. OpenAI also paired the launch with a limit reshuffle. In its forum post, the company said the $100 plan includes all existing Pro features plus 5 times more Codex usage than Plus, and said Plus usage is being “rebalanced” to support more sessions across the week instead of longer bursts in one day. For a limited-time launch push through May 31, OpenAI said $100 Pro subscribers can get up to 10 times the Codex usage of ChatGPT Plus. That kind of temporary boost is a classic way to get power users to move habits, because developers notice friction fastest when a tool cuts them off mid-workflow. The timing also lines up with a model upgrade. OpenAI introduced GPT-5.4 in March and said the model folds in the coding strengths of GPT-5.3-Codex while becoming the default choice for broad work across ChatGPT, Codex, and the application programming interface. That matters because bigger coding sessions eat context, which is the model’s working memory for the text it can keep in view at once. OpenAI’s application programming interface docs say GPT-5.4 and GPT-5.4 pro support a 1.05 million token context window, which is large enough to hold very long codebases, logs, and documentation in one session. The competitive target is not subtle. CNBC and other outlets framed the move against Anthropic’s Claude Code, because both companies are trying to own the market for developers who now treat artificial intelligence less like autocomplete and more like a junior engineer that never sleeps. So the real news is not just a cheaper subscription tier. OpenAI is turning coding into its own pricing lane, with limits, rate cards, specialized models, and plan tiers that look more like cloud software bills than the old flat-fee chatbot subscriptions people got used to in 2023.