OpenAI launches $500 Pro 500 tier and cuts limits on $200 Pro plan

OpenAI now sells a $500-per-month subscription. The new Pro 500 plan, announced at the company's DevDay event, offers OpenAI's highest usage allowance and includes access to a new "Ultrafast" feature — a price point that resets the ceiling of what a consumer-and-developer AI subscription can cost.

The move matters beyond OpenAI's own pricing page. At the end of 2024, OpenAI became the first AI lab to introduce a $200 Pro plan, and nearly every other provider — including Anthropic and Perplexity — later announced their own $200 tiers. If that pattern holds, the $500 Pro 500 plan could become a template that competitors copy, pushing the top end of the AI subscription market sharply upward.

What $500 per month buys

The Pro 500 plan bundles two things: OpenAI's highest usage allowance and the new "Ultrafast" speed tier. "Ultrafast is our premium speed tier for workloads where speed matters most," the company explains. With the feature enabled, OpenAI's Codex and Work apps will generate up to 300 tokens per second, which the company says makes them capable of writing code and carrying out other tasks faster. The Ultrafast speed tier is also available to API users, extending the offering beyond subscription customers to developers building on OpenAI's platform.

For context on the arithmetic: the Pro 500 plan costs two and a half times what the existing $200 Pro plan costs. OpenAI is framing it as the option for users whose workloads genuinely depend on throughput and volume — the developers running long coding sessions in Codex, or teams leaning on the Work app for high-volume tasks, where waiting on token generation translates directly into lost time.

The $200 tier gets smaller

OpenAI did not just add a more expensive plan. It made the existing $200 Pro plan less appealing at the same time.

In Codex and Work, $200 Pro subscribers will see their included usage cut in half — from 20x of what OpenAI offers to Plus users, down to 10x of that same allowance. In ChatGPT, GPT-6 Pro message caps will decrease from 200 to 100 per week.

The company is softening the landing for current customers. According to OpenAI, existing subscribers will keep their current limits for a time, and will later receive a one-time credit to help them make the most of their new reduced allowances. That structure — temporary grandfathering plus a credit rather than a permanent exemption — signals that OpenAI views the reduced limits as the new baseline for the $200 tier, not a temporary adjustment.

The combined effect is clear. The gap between the $200 plan and the $500 plan is now wider in both directions: the cheaper tier holds less than it did yesterday, and the premium tier holds more and runs faster. Users who found the old 20x Pro allowance sufficient now face a choice between accepting half of it or more than doubling their monthly spend.

Dots: personal AI agents that don't count against your cap

All Pro subscribers will get a single Dot — the name OpenAI is giving to its new personal AI agents. The pricing detail that will likely draw the most attention: any conversations a user has with their Dot won't count toward their ChatGPT usage limit.

That exemption is notable in a release that otherwise tightens usage caps. It suggests OpenAI wants adoption of Dots to be frictionless even for subscribers whose allowances just shrank, positioning the agent product as a headline feature rather than another metered service.

Sign in with ChatGPT reaches 16 partners

OpenAI also announced a distribution play for its models outside its own apps. The company is partnering with 16 companies, including Notion and Cognition, to let Plus and Pro customers run OpenAI models in those services using the new "Sign in with ChatGPT" option. There is no need to first set up an API key.

The removal of the API-key step lowers the barrier for non-developer subscribers. A Plus customer who already pays OpenAI monthly can now carry their entitlement into third-party tools, which extends the value of an OpenAI subscription without requiring any technical setup — and puts OpenAI's billing relationship, rather than the partner's, at the center of those integrations.

GPT-6.1 Sol: frontier-class performance at one fifth the price

The third major announcement is a new model: GPT-6.1 Sol. The company says the new model offers nearly the same performance as its current frontier system, GPT-6.0 Astra, in areas like coding and computer use, while costing one fifth of the price.

That pricing claim is the story's research-and-market stakes in miniature. If a model at 80 percent below frontier cost delivers near-frontier results on the workloads developers actually run — coding and computer use — the pressure on competing providers to justify premium model pricing intensifies. OpenAI is effectively splitting its lineup: Astra as the flagship, Sol as the efficient workhorse.

OpenAI also claims the new model is safer to use, and it put that claim in specific terms. "GPT-6.1 Sol is more transparent about its limitations and more reliable at respecting user intent and safety constraints," the company says. "In challenging evaluations, it shows lower failure rates than GPT-6 Sol on transparency about broken search tools, respecting explicit restrictions, and avoiding unauthorized outcomes during agentic tasks."

The mention of agentic tasks ties back to the Dots announcement. As OpenAI ships personal AI agents that act on users' behalf, the failure mode of "unauthorized outcomes" — an agent doing something it was not supposed to do — becomes a first-order safety concern rather than an edge case. OpenAI is arguing that Sol was evaluated specifically against that risk and performed better than its predecessor.

Availability

Plus, Pro, Business, Enterprise and Education users can start using GPT-6.1 Sol in Codex and Work starting today, with availability in ChatGPT to follow at a later date. The staggered rollout puts the model first in the hands of OpenAI's most active tool users — developers and work app customers — before the general ChatGPT audience.

Why this DevDay adds up to a pricing reset

Taken together, the announcements form a coherent strategy. OpenAI built a new top tier at $500 with exclusive speed, cut the value of the $200 tier to widen the upgrade incentive, threw in an unmetered agent to sweeten both Pro plans, extended subscriptions into 16 third-party services, and shipped a cheaper near-frontier model that lowers the cost of the heavy usage the premium tiers enable.

The precedent argument cuts both ways. Competitors matched the $200 tier after OpenAI introduced it in late 2024; whether they follow to $500 will test how much the market's top spenders will bear. For existing $200 Pro subscribers, the immediate question is simpler: their weekly GPT-6 Pro message cap is heading from 200 to 100, and their Codex and Work allowance from 20x to 10x the Plus baseline — the one-time credit notwithstanding. The next signal to watch is whether Anthropic, Perplexity and others match the new ceiling, or position their $200 tiers against OpenAI's diminished offering instead.