OpenAI reversed a planned usage reduction on its ChatGPT Pro 200 plan after subscriber pushback and instead doubled Pro quotas. Pro 100 and Pro 500 now receive twice their previous Work and Codex allowances, and Pro 200 doubled its lower quota, in a direct price-per-value fight with Anthropic just as the rival prepares its public offering.
A rare public U-turn
Subscription pricing changes are usually announced, defended, and quietly absorbed. OpenAI did the opposite. It floated a reduction in usage limits on the Pro 200 tier, users pushed back loudly, and the company not only walked it back but flipped it into a raise. The new limits more than compensate: Pro 100 and Pro 500 get double their earlier Work and Codex quotas, and existing customers receive the increase automatically while new subscribers get it from day one. The Plus plan is unchanged.
It is a small, human moment in an industry that often communicates through press releases and benchmark charts. A company with a 1.4 trillion dollar private valuation actually listened to its paying users and gave them more. That is the fun part. The whiplash is the point: the messaging went from "we are tightening" to "we are doubling" in the span of a single news cycle.
Why now, and why this matters
OpenAI's own framing points to efficiency. Its recent GPT-6.1 models, Terra and Sol, do more per token, so the company says it can push usable capacity higher without changing the sticker price. That is a real engineering story: when inference gets cheaper, the same subscription can quietly buy more. But the timing is not accidental. The Decoder and others read the move as part of a direct contest with Anthropic for the business market, exactly as Anthropic lines up an IPO that could value it at up to 2 trillion dollars.
When two frontier labs fight over who gives the subscriber more for the same money, customers win in the short run. The risk is that "more for the same" becomes a race to the bottom on margins that neither can actually sustain once the compute bills arrive. OpenAI was careful to note that doubling a quota does not mean twice as many completed tasks, because consumption depends on the model, the task type, and the settings. Translation: your mileage will vary, and the bill for all this generosity is paid somewhere in the data center.
The competitive choreography
Think about the choreography of late October 2026. Google pushes Gemini 4 Carbon to the top of the intelligence index. Anthropic files for an IPO that turns it into a public company whether the market likes it or not. OpenAI answers with a generosity move on pricing rather than a model release. Each lab is competing on a different axis — capability, capital, and now customer value — and the user is the battlefield.
There is also a quiet signal about where OpenAI's growth comes from. The Pro tiers are the power-user and developer segments, the same people who run Codex agents and long-horizon work. Doubling their quotas is a bet that heavy users are the ones worth retaining, and that the best defense against a newly public, well-funded competitor is a fanatically satisfied base of builders.
Key data points
Change: planned Pro 200 limit cut reversed and converted into a quota increase.
New limits: Pro 100 and Pro 500 get twice their previous Work and Codex allowances; Pro 200 doubled its lower quota.
Rollout: automatic for existing customers; immediate for new subscribers. Plus plan unchanged.
Stated driver: efficiency gains from GPT-6.1 models Terra and Sol.
Competitive context: read as a price-per-value fight with Anthropic ahead of its IPO.
Caveat from OpenAI: doubling a quota does not equal twice as many completed tasks.
Sources: dailyaithread.com, The Decoder, aibriefing.dev, threerouter
Top comments (0)