How much is ChatGPT Pro?
ChatGPT Pro is sold as three monthly plans — Pro 100 at $100, Pro 200 at $200 and Pro 500 at $500. All three include Pro models, Codex, deep research, image creation, memory and file uploads. The one feature OpenAI reserves for a single tier is GPT-6 Astra Ultrafast, on Pro 500 only; everything else between them is allowance — and OpenAI no longer says how much more than Plus any tier includes. None is unlimited, and none bills annually.
- Pro 100
- $100/mo lowest Pro allowance
- Pro 200
- $200/mo smaller for new subscribers
- Pro 500
- $500/mo only tier with Ultrafast
- vs Plus
- $20/mo the tier below
- Multiple of Plus
- Not stated 5× / 20× wording withdrawn
- Annual option
- None monthly billing only
- Unlimited
- No per-model allowances apply
- Verified
- 2026-10-02 vendor plan sheet
Three prices, one extra feature — the rest is allowance OpenAI no longer measures for you.
OpenAI's Pro article now lists three monthly plans in a table with two columns that matter: price and Ultrafast. Pro 100 and Pro 200 read Not included; Pro 500 reads Included. The rest of the comparison is two sentences: "Pro 200 includes more usage than Pro 100. Pro 500 offers the highest included usage of the three plans." That is an ordering, not a quantity. The "5× and 20× Plus usage" figures this page used to quote are gone from the article, from OpenAI's pricing page and from its Codex pricing docs.
Pro 200 also changed without changing its name. New subscriptions that are not grandfathered get a lower usage allowance than Pro 200 used to include, at the same $200. Anyone with an active Pro 200 subscription between September 22 and 10 a.m. Pacific on September 29, 2026 keeps the previous allowance through October 29, 2026, then moves to the lower one. OpenAI's stated reason is its "increasingly efficient models"; it publishes no figure for either allowance.
Below the plan names, the differences are in how allowances behave. In Chat, GPT-6 Pro — powered by GPT-6 Astra — carries its own limits: on Pro 200, reaching the weekly GPT-6 Pro limit switches you to GPT-5.6 Thinking at Medium, while on Pro 100 GPT-6 Pro and GPT-5.6 Sol Pro draw from one shared weekly allowance, so switching between them buys nothing. In ChatGPT Work and Codex, all three Pro plans currently have no five-hour limit, only the allowance itself. Ultrafast spends that allowance at 8× the Standard rate; Fast mode, on any plan that has it, at 2.5×.
The ceiling is real on every tier. At a model's limit it is unavailable until the allowance resets; OpenAI says you cannot bypass a limit or increase it through a setting, and that Support cannot reset ChatGPT or Codex limits. What it sells instead is a way past the edge for some features: credits for Codex, ChatGPT Work and the Office add-ins once the included allowance is spent, and a paid instant reset of the Work and Codex allowance. Neither touches the Chat model limits. Prices here come from our verified plan sheet, checked 2026-10-02, so this page cannot disagree with the subscription-vs-API calculator.
Whether the jump from Plus is worth it.
A meaningful share of searches for this price use the word professional. There is no ChatGPT Professional tier — the consumer ladder is Free, Go, Plus and Pro (now in three sizes), with Business, Enterprise and Edu for organisations. The name persists from early coverage of the paid tier and from other vendors who do use it.
If you were sent an invoice for "ChatGPT Professional", check who actually charged you; resellers repackaging API access under invented plan names are a recurring pattern in this market.
Until late September 2026 the two Pro prices came with a multiple of Plus attached. Today no OpenAI page we read puts a number on any Pro allowance, and the one place Pro used to be measured is now empty: OpenAI publishes five-hour message estimates for Plus in Work and Codex, but Pro plans have no five-hour window to estimate.
So the honest way to size Pro is your own meter. Settings → Usage shows the remaining allowance and reset times. Start at Pro 100, watch it through a normal working week, and move up only if it runs dry — the plans differ in allowance, and in Ultrafast, not in which models you can open.
OpenAI's Pro article used to spell out that an upgrade applies at once and restarts the billing cycle, while a downgrade waits for renewal. The current version says only that Settings → My Plan shows your subscription and the plan changes available to it. What is still documented:
- Cancel — renewals stop and you keep the plan to the end of the period you paid for.
- Drop to Go — your current plan runs to the end of the billing cycle, then switches; no refund for the difference.
- Billing date — a monthly subscription renews on the day you subscribed, and OpenAI does not move invoice dates.
How a mid-cycle move between Pro tiers is prorated is not stated anywhere we found. Read the amount the checkout shows before confirming.
Pro earns its price in two situations: you hit Plus limits often enough that the interruption costs you more than the difference, or you need the Pro reasoning models — GPT-6 Pro and GPT-5.6 Sol Pro — that Plus cannot select in Chat. Both are real for people running long coding sessions, heavy deep-research workloads or all-day agentic work, and neither describes most people asking this question.
Before buying headroom, check whether the workload belongs on a subscription at all. Sustained programmatic volume is usually cheaper on per-token API rates, and our subscription-vs-API calculator settles it with your own numbers instead of a rule of thumb. For the tier-by-tier view, we have Plus vs Pro written out.
Three cheaper answers to "I keep hitting limits".
Pro is the right purchase for sustained heavy use and an expensive one for everything else. Prices come from our verified plan sheet, checked 2026-10-02.
Frequently asked.
Quick follow-ups people search after this question.