OpenAI To Decrease Limits On Pro Plan From 20 Times To 10 Times Of Plus Plan

OpenAI’s Tibo Sottiaux had previously Saif that the company’s $200 plan would see its usage slashed in half for customers, and now there’s more clarity on what that’ll look like.

OpenAI is cutting the usage included in its $200 Pro plan for existing subscribers. Starting October 30, 2026, included usage in ChatGPT Work and Codex falls from 20 times to 10 times the ChatGPT Plus allowance. The weekly limit for GPT-6 Pro in Chat also drops from 200 to 100 messages. Subscribers keep their current limits through October 29.

The cut follows a change OpenAI announced for new subscribers earlier this week. The company had paused sign-ups for the $200 tier on September 10, citing overwhelming demand for GPT-6 Astra, which put heavy strain on its compute. It reopened the plan on September 30, but with roughly half the usage it offered before. OpenAI’s Thibault Sottiaux said the new calculation nets out to half the dollar value in API spend compared with the old plan. The latest notice extends the same reduction to people already on the plan, after a grace period.

A new $500 plan

Alongside the cuts, OpenAI has introduced a new top-tier Pro 500 plan at $500 a month. It offers the highest usage limits of any individual plan, at 25 times the Plus allowance, according to OpenAI’s help documentation. It is also the only Pro plan that includes “Ultrafast”, a speed tier that OpenAI says delivers up to 8x faster performance in Codex. The plan was announced at the company’s DevDay event.

OpenAI’s Pro lineup now has three tiers: $100 at 5x Plus, $200 at 10x Plus for new subscribers, and $500 at 25x Plus. That works out to about $20 per Plus-equivalent unit at every level, so the bigger plans no longer carry a volume discount.

OpenAI has also said it will not bring back the five-hour usage window on the $200 plan. Users can spend their weekly allowance whenever they like, which matters for heavy Codex users who work in long sessions.

Compute crunch

The changes point to the pressure OpenAI is under on compute. Demand for its latest models has outstripped capacity, and the company is rationing access through plan design rather than only through rate limits. That is happening as competitors like Anthropic keep shipping new models aimed at the same coding and agentic workloads.

For subscribers, the practical effect is that the $200 plan now buys less than it did a month ago. Those who need the old headroom will have to pay for the $500 tier.

Posted in AI