If Codex is becoming your pair programmer, you have a clear upgrade path
In the quiet arithmetic of subscription tiers, OpenAI has drawn a new line between the casual and the committed — a $100-per-month plan for developers who have come to rely on AI not as a novelty, but as a working partner. Positioned between its $20 and $200 offerings, the new tier mirrors Anthropic's own pricing architecture almost exactly, suggesting that the real competition in AI is no longer about capability alone, but about who can hold a developer's loyalty at scale. A promotional window offering double the usual usage boost through May signals that OpenAI is not merely setting a price — it is making an argument.
- The AI coding market is stratifying fast, and OpenAI is drawing clear lines between hobbyists, daily users, and serious builders.
- The new $100 tier lands almost perfectly on top of Anthropic's Max 5x plan, turning subscription pricing into a direct competitive battleground.
- A limited-time promotion offering 10x usage through May 31 creates urgency — and a calculated bet that developers who taste more will pay to keep it.
- OpenAI is also quietly rebalancing Plus usage to prevent single-day burnout, nudging power users toward the Pro upgrade rather than letting them game the lower tier.
- The three-tier architecture now gives OpenAI a clear answer for every developer: experiment free, integrate at $20, or build seriously at $100 and beyond.
OpenAI has introduced a $100-per-month subscription tier aimed at developers who rely on Codex — its AI coding agent — for intensive, sustained work. The new plan sits between the existing $20 Plus and $200 Pro offerings, with each tier defined by how much Codex usage it unlocks: baseline, five times, and twenty times, respectively.
The pricing structure is a near-exact mirror of Anthropic's own tiered model for Claude Code, and the resemblance is unlikely to be coincidental. Where the $20 Plus plan suits steady, day-to-day coding assistance, the new $100 tier is designed for developers who need longer, deeper sessions — the kind of work that unfolds over hours rather than quick queries.
To accelerate adoption, OpenAI is offering a promotional 10x usage multiplier for anyone who subscribes before May 31 — double the plan's standard allowance — a clear attempt to pull developers into the higher tier and keep them there once the promotion expires.
Alongside the new tier, OpenAI is adjusting how Plus subscribers access Codex, redistributing their monthly allowance more evenly across the week rather than allowing it to be consumed in a single session. The change nudges heavy users toward upgrading while smoothing the experience for everyone else.
Taken together, the moves reflect a maturing market in which AI coding tools are no longer a curiosity but a professional dependency — and where the question is no longer whether developers will pay, but how much, and for what.
OpenAI has introduced a new subscription tier aimed squarely at developers who spend their days writing code with artificial intelligence. The $100-per-month ChatGPT Pro plan sits between the company's existing $20 Plus offering and a $200 Pro tier, each tier distinguished by how much you can use Codex—OpenAI's AI coding agent built into ChatGPT.
Codex is the engine that writes code, debugs, and refactors alongside you in the chat interface. It competes directly with Anthropic's Claude Code, and the new pricing structure suggests OpenAI is watching Anthropic's moves closely. Where Anthropic offers a $20 Pro plan, a $100 Max 5x plan, and a $200 Max 20x plan, OpenAI now mirrors that architecture almost exactly.
The math is straightforward. The $20 Plus plan gives you a baseline amount of Codex usage. The new $100 tier provides five times that allowance. The $200 tier goes to twenty times. OpenAI frames the $100 plan as the right choice for developers who need longer, more intensive coding sessions—the kind of work that might stretch across hours of back-and-forth with the AI. The $20 plan, by contrast, is positioned as steady, day-to-day usage: you get consistent access throughout the week rather than the ability to burn through your quota in a single marathon session.
To sweeten the launch, OpenAI is running a limited promotion through May 31. Anyone who signs up for the $100 plan during this window gets ten times the Codex usage of the $20 Plus tier—a significant bump that undercuts the usual five-times multiplier. It's a clear signal that OpenAI wants developers to try the higher tier and, presumably, stay there once the promotion ends.
The company is also making a smaller adjustment to how Plus subscribers experience Codex. Rather than allowing them to exhaust their monthly allowance in a single day of heavy use, OpenAI is rebalancing the system to distribute usage more evenly across the week. This change smooths out the experience for casual users while making the case for upgrading to Pro if you need sustained, intensive coding work.
The move reflects a broader shift in how AI companies are monetizing their products. As coding tools become more central to how developers work, the market is stratifying. You can use the free tier or the basic subscription if you're experimenting. You can use Plus if you're integrating AI into your daily workflow. And now, if you're building something serious—if Codex is becoming your pair programmer—you have a clear upgrade path to Pro. OpenAI is betting that enough developers will find that path worth taking.
Citações Notáveis
The $100 plan is best for longer, high-effort Codex sessions, while Plus remains the choice for steady, day-to-day usage— OpenAI
A Conversa do Hearth Outra perspectiva sobre a história
Why does OpenAI need a middle tier? Why not just have Plus and the $200 plan?
Because developers don't all have the same needs. Some use Codex casually. Others live in it. A $20 plan might feel too constrained for serious work, but jumping straight to $200 feels like overkill. The $100 tier is the Goldilocks option.
And the promotion—ten times usage through May 31—that's aggressive, isn't it?
Very. It's a customer acquisition play. OpenAI is saying: try this tier, get a taste of what real Codex power feels like, and by the time the promotion ends, you'll be too invested to leave.
What about the rebalancing of Plus usage? That sounds like a nerf.
It is, but it's a strategic one. They're saying: if you want to binge-code with Codex all day, you need to pay more. If you want steady, reliable access, Plus is fine. It pushes power users toward Pro without making Plus feel broken.
Anthropic has the exact same pricing structure. Is OpenAI copying them?
Or they're both converging on what the market will bear. When your competitor has three tiers at $20, $100, and $200, and they're working, you don't reinvent the wheel. You match it and compete on the product itself.
So this is really about the coding market heating up?
Exactly. Coding is where AI adds the most immediate, measurable value. Developers will pay for it if it saves them time. Both OpenAI and Anthropic know that. The pricing tiers are just the two companies drawing lines around who gets what level of that value.