OpenAI tests a $100 Codex tier
OpenAI added a $100/month Pro subscription for Codex users that sits between existing tiers and offers roughly five times the usage of the $20 plan, with a temporary tenfold boost through May 31. The change reads like an experiment to capture heavy‑use developers who balk at steep jumps between hobby and enterprise pricing. That kind of mid‑market packaging experiment matters because similar tier gaps will appear for autonomy and task‑volume on consumer agent platforms. (cryptobriefing.com)
OpenAI just inserted a new price point into ChatGPT: $100 a month for people who use Codex heavily, instead of forcing them to choose between $20 Plus and the older $200 Pro tier. OpenAI said the new plan went live on April 9 and gives about five times the Codex usage of Plus. (community.openai.com) Codex is OpenAI’s coding agent inside ChatGPT, which means it does more than spit out snippets. It can work through longer programming jobs, review code, and handle cloud tasks that burn through usage limits faster than casual chat does. (developers.openai.com) That usage meter is not a simple message counter. OpenAI says a small script can use only a fraction of an allowance, while a big codebase or a long task that needs more context can consume much more. (developers.openai.com) The old jump was steep. A developer who outgrew the $20 plan had to leap straight to $200, which is the kind of pricing cliff that makes people ration usage instead of building around the tool. (techcrunch.com) OpenAI is sweetening the launch with a temporary boost through May 31, 2026. During that window, the new $100 tier gets up to ten times the Codex usage of Plus, and OpenAI’s announcement says the highest limits shown in ChatGPT are temporarily elevated. (community.openai.com) (venturebeat.com) The company did not scrap the top tier. The $200 Pro option still exists, and OpenAI’s community post says it keeps “significantly more fire power” than the new $100 plan after the promotional period ends. (community.openai.com) This is aimed at a specific kind of buyer: the person who codes enough to hit limits every week but not enough to justify an enterprise contract. TechCrunch reported that OpenAI framed the move directly against Anthropic, which has long sold a $100 Claude plan to developers. (techcrunch.com) OpenAI is also adjusting around Codex outside consumer subscriptions. Its developer pricing page says Business and Enterprise customers can buy extra credits, which turns heavy coding use into something closer to metered cloud spending than a flat all-you-can-eat seat. (developers.openai.com) That makes the new $100 plan look less like a random discount and more like packaging. OpenAI now has a ladder where Plus covers light use, $100 catches serious individual developers, and $200 or business plans absorb people running bigger parallel workflows. (community.openai.com) (developers.openai.com) If this tier sticks, it is a clue about how artificial intelligence agents will be sold next. Once products start charging for longer tasks, parallel jobs, and autonomy instead of just chat access, the biggest fight will be over the middle band where people want much more than hobby limits but still flinch at enterprise prices. (techcrunch.com) (developers.openai.com)
