OpenAI Launches $100/Month ChatGPT Pro Tier for Codex Power Users

The $100 tier was not the product of a deliberate go-to-market strategy. It was a response to a demand signal that OpenAI could no longer ignore. Codex went from 2 million to 3 million weekly users in just seven days — between March 31 and April 7, 2026 — while already sitting at 5x growth over three months. That kind of acceleration creates a specific infrastructure and monetization problem: existing Plus tier limits were being hit constantly by the heaviest users, and those users were the most valuable ones to retain.

Sam Altman’s own framing — “by very popular demand” — is unusual for a major product launch. CEOs typically frame new tiers as strategic expansions. Altman’s language instead acknowledged that the user base had outgrown the existing menu. The launch-promotion decision to offer 10x Codex usage through May 31 (rather than the standard 5x) reinforces this reading: OpenAI needed to move fast, reward early adopters, and establish habitual usage before the promotional window closed. The tier wasn’t scheduled — it was summoned by usage data.

The most underreported fact in the launch coverage is that Anthropic already offered a $100/month Claude Max 5x plan before OpenAI announced this tier. The full pricing ladders are now nearly identical: both companies offer $20/month, $100/month, and $200/month personal tiers, with usage multipliers at each step (5x and 20x versus Plus/Pro equivalents). OpenAI did not invent this structure — it matched it.

This matters because it reframes the competitive narrative. OpenAI is often cast as the incumbent and Anthropic as the challenger. In the developer-tools segment, that dynamic appears reversed. Anthropic had Claude Code, a $100 tier, and a developer-first positioning before OpenAI formalized its equivalent. The YouTube commentary — with titles like “OpenAI Just Killed Claude Code” — reflects a media instinct to declare a winner, but the structural reality is that OpenAI spent this launch playing catch-up on pricing architecture. The real competitive question is not which company launched first, but which one can sustain the infrastructure reliability and model quality that justify $100/month for professional daily use.

Buried in the announcement is a change that received almost no media attention: OpenAI simultaneously rebalanced Codex usage limits for existing Plus subscribers. The rebalancing shifted Plus toward “more consistent daily sessions” rather than single long sessions. In practice, this means users who relied on Plus for extended, uninterrupted Codex work — the exact behavior that defines serious development workflows — lost capacity they previously had.

This is a structurally significant move. It creates pressure on the most active Plus users to upgrade to the $100 tier, which is precisely the conversion OpenAI needs to justify the new tier’s existence. The rebalancing was not framed as a cut — it was framed as a quality-of-service improvement. But the net effect for any developer who regularly ran long Codex sessions on Plus is a degraded experience unless they pay five times more per month. This kind of silent repricing of existing tiers is worth watching as a pattern: as AI companies face infrastructure costs from power users, they will increasingly restructure limits to route those users toward higher-priced tiers rather than absorb the compute cost within existing plans.

Both OpenAI and Anthropic have converged on the same insight: the highest-value personal subscription customers are not casual chatbot users — they are developers who integrate AI into their daily professional output. The $100 tier is not primarily about ChatGPT the consumer product; it is about Codex the software engineering agent. OpenAI’s messaging consistently emphasizes “longer, high-effort Codex sessions” and frames the tier as being for people whose “professional output depends” on consistent AI performance.

The second-order implication is that developer tooling — historically monetized through API usage and enterprise contracts — is now being pulled into the subscription model. Subscriptions are stickier than API billing, generate predictable revenue, and create habitual daily engagement that makes switching costly. By creating a $100 developer-adjacent subscription tier, both OpenAI and Anthropic are betting that the next large revenue pool is individual developers paying monthly, not just enterprises paying annually. The competitive intensity at the $100 price point — where both companies now have matching offers — suggests this market segment is seen as large enough to fight over directly, and that the winner will be determined by model quality and infrastructure reliability rather than price alone.