ChatGPT Pro usage limits in Codex: the Pro 200 cut was canceled (October 2026)
As of 2026-11-06: OpenAI canceled the Pro 200 reduction on 2026-10-29, and Pro 200 subscribers still on the lower allowance get twice their previous allowance from that day, bringing them up to the same limit; Pro 100 and Pro 500 also get twice their previous Work and Codex allowance. The 2026-09-29 email had announced that Pro 200 would move to a lower allowance; under the 10-29 update, subscriptions that already had the higher allowance keep it. Among Pro plans, Pro 500 is the only one with Ultrafast (supporting GPT-6 Astra and GPT-6.1 Sol), which uses included limits at 8x the Standard rate and is billed at 6x on purchased credits. From December, ChatGPT and Codex will share one allowance; OpenAI has not published an exact date.
Updated 2026-11-06
One key for Claude, GPT and leading Chinese models, billed per token, up and running in minutes.
- Billed per token — live rates on /models
- Pay by card (Visa / Mastercard), Apple Pay, Google Pay or crypto
- Self-service: top up and activate instantly
Four official numbers
The day the Pro 200 cut was canceled
On 2026-10-29 OpenAI canceled the Pro 200 reduction announced in its 2026-09-29 email: subscriptions that already had the higher allowance keep it, and Pro 200 subscribers on the lower allowance were doubled to the same limit from that day.
Where the three Pro plans stand
Since 2026-10-29, Pro 100 and Pro 500 include twice their previous Work and Codex allowance, and Pro 200 on the lower allowance is brought up to the same limit as the higher one. All three Pro plans currently have no five-hour limit in Work and Codex, but they still have an included allowance.
Ultrafast on included limits
Compared with Standard mode for the same model, Ultrafast for GPT-6 Astra and GPT-6.1 Sol uses included limits at 8x the rate; Fast mode uses 2.5x. OpenAI says these are consumption rates, not speed claims.
Ultrafast on purchased credits
With purchased credits, Ultrafast is billed at 6x the Standard rate and Fast at 2x. On Pro 500, included usage is used first, then credits.
What actually changed in October
As of 2026-11-06, the OpenAI Help Center update of 2026-10-29 says two things. First, Pro 100 and Pro 500 now include twice their previous Work and Codex allowance. Second, the Pro 200 reduction announced in the 2026-09-29 email was canceled: subscriptions that already had the higher allowance keep it, and Pro 200 subscribers on the lower allowance get twice their previous allowance from 2026-10-29, bringing them to the same limit.
In other words, the move to a lower allowance after 10-29 that the email announced will not happen. Pro 100 and Pro 200 do not include Ultrafast; only Pro 500 does. ChatGPT Work and Codex share the same allowance.
What the official text says
The Help Center sentence of the 2026-10-29 update reads: "We’ve also canceled the Pro 200 reduction announced in our September 29 email." It continues: subscriptions that already had the higher allowance keep it, and Pro 200 subscribers on the lower allowance get twice their previous allowance from 10-29, bringing them to the same limit; the same update says Pro 100 and Pro 500 now include twice their previous Work and Codex allowance.
The Help Center's earlier statement that Pro 200 moves to a lower allowance after 10-29 has been withdrawn by this update. The official DevDay 2026 page (2026-09-29) presents Astra Ultrafast on the Pro 500 plan. One more announcement: from December, Plus, Pro and Business allowances will be shared across ChatGPT and Codex and measured in tokens, and Instant chat does not count; OpenAI has not published an exact date or the pool size for each plan.
Timeline
According to the official Help Center page as archived on 2026-09-29: as of 2026-09-10, new sign-ups and upgrades to ChatGPT Pro 200 were paused; existing Pro 200 subscriptions were not affected.
DevDay 2026: the official page presents GPT-6 Astra Ultrafast in Codex and ChatGPT Work on the Pro 500 plan. An email the same day announced a lower allowance for Pro 200; that reduction was canceled on 10-29.
OpenAI cancels the Pro 200 reduction: subscriptions with the higher allowance keep it, Pro 200 on the lower allowance is doubled to the same limit, and Pro 100 and Pro 500 get twice their previous Work and Codex allowance.
Confirmed vs not published
Confirmed (verbatim on official pages)
2026-10-29: Pro 100 and Pro 500 get twice their previous Work and Codex allowance; the Pro 200 reduction announced in the 2026-09-29 email is canceled, and Pro 200 on the lower allowance is doubled to match from that day. Pro 100 / 200 / 500 currently have no five-hour limit in Work and Codex but still have an included allowance; the same page notes weekly limits may also apply, and the allowance does not correspond to a fixed number of messages or tasks.
Among Pro plans only Pro 500 includes Ultrafast (supporting GPT-6 Astra and GPT-6.1 Sol); Plus, Pro 100, Pro 200 and Business do not include it at launch, even with purchased credits. Ultrafast uses included limits at 8x and purchased credits at 6x; Fast uses 2.5x and 2x.
Not published by OpenAI
OpenAI has not published the multiple of Plus for Pro 100 / 200 / 500 or exact weekly numbers. It has also not published the pool size per plan after the December shared allowance, or the exact start date. Your own usage dashboard is the reference.
Plan allowance or API key
Plan allowance (Pro 100 / 200 / 500)
Good for interactive coding and cloud features such as Codex cloud tasks and GitHub code review. Usage depends on model, context, reasoning and speed mode, and you can buy credits when it runs out. Among Pro plans, Ultrafast requires Pro 500.
API key (API pricing)
The official pricing page says that with an API key you "Pay for Codex usage based on API pricing", and that all users may also run extra local chats using an API key, charged at standard API rates. The trade-off: no cloud features such as GitHub code review or Slack. OpenAI describes it as great for automation in shared environments like CI.
What to do when you hit the Codex limit
- 1Open the usage dashboard, or type /status in a Codex CLI session, to see your remaining limits and reset time.
- 2The ways out OpenAI lists: wait for the reset, use a saved reset, or continue with eligible credits, depending on your account; switching models does not restore allowance in a shared usage pool.
- 3Move CI and batch scripts that do not need cloud features to an API key: the official pricing page says API-key usage is paid at API pricing and describes it as great for automation in shared environments like CI.
On QCode
QCode sells API access, not ChatGPT subscriptions, billed per token; per-model prices are listed on /models. With a QCode key in Codex CLI you can call gpt-6.1-sol, gpt-6-sol and gpt-5.6-sol; the setup steps are in our Codex CLI third-party API guide. Ultrafast is a ChatGPT subscription feature (among Pro plans, only on Pro 500) and is not part of the QCode API.
FAQ
Will the Pro 200 allowance in Codex go down?
No. On 2026-10-29 OpenAI canceled the Pro 200 reduction announced in its 2026-09-29 email: subscriptions that already had the higher allowance keep it, and Pro 200 on the lower allowance is doubled to the same limit from that day.
What multiple of Plus is the Pro 200 allowance?
OpenAI has not published a multiple. It only says Pro 100 and Pro 500 now include twice their previous Work and Codex allowance, and that the allowance does not correspond to a fixed number of messages or tasks. Check your usage dashboard for real numbers.
What happens to Pro 200 subscribers who were on the lower allowance?
OpenAI says that starting 2026-10-29, Pro 200 subscribers on the lower allowance get twice their previous Work and Codex allowance, bringing them up to the same limit as the higher one.
Does Pro 200 include Ultrafast?
No. OpenAI says Plus, Pro 100, Pro 200 and Business do not include Ultrafast at launch, even with purchased credits. Among Pro plans only Pro 500 does, and included usage goes first, then credits.
What is the difference between 8x and 6x for Astra Ultrafast?
8x is the rate at which it uses included subscription limits; 6x is the billing rate on purchased credits. Both are relative to Standard mode for the same model and apply to Ultrafast on both GPT-6 Astra and GPT-6.1 Sol, and OpenAI says they do not describe speed. On Pro 500, included usage goes first, then credits.
What can I do when the Codex limit runs out?
OpenAI says you can wait for the reset, use a saved reset, or continue with eligible credits, depending on your account; eligible personal Plus and Pro accounts can also buy an instant reset, depending on account and billing country; switching models does not restore allowance in a shared usage pool. After ChatGPT and Codex share one allowance in December, the options listed are to wait for the reset, upgrade your plan, or buy credits. You can also move part of the work to token-billed API usage.
Sources
OpenAI Help Center: the 2026-10-29 entry in the ChatGPT Business release notes, Understanding ChatGPT and Codex usage limits, and Managing usage with GPT-6 Astra in Work and Codex, all read in a browser on 2026-11-06; the Codex Pricing and Speed pages on developers.openai.com, fetched on 2026-11-06. The Pro 200 reduction described in the versions of this page from 2026-10-07 and 10-14 was canceled by OpenAI's 2026-10-29 update.
Keep part of your Codex work off the plan allowance
One key for gpt-6.1-sol, gpt-6-sol and gpt-5.6-sol, billed per token. Per-model prices are on /models.
Related reading
The Codex weekly allowance vanished in hours
Where the allowance goes, how reasoning levels differ, and what to do.
When ChatGPT Pro is unavailable, use the API path
The official wording during the September Pro 200 pause, and the API route.
Codex CLI Third-Party API Setup Guide
How to point Codex CLI at a third-party API.
Crypto payment guide
Nine coins, 56 networks, zero platform fee for AI coding API credits.
Figures and quotes checked on 2026-11-06; the official OpenAI pages prevail, and OpenAI may change them without notice. Your usage dashboard is the reference for plan limits; model availability is listed on /models.