ChatGPT Pro 200 drops to 10x Plus usage after October 29: what Codex users should check before choosing Pro 100, 200 or 500

If you pay $200 a month for ChatGPT Pro mainly to run Codex, the date to watch is October 29, 2026, not November 1. That is the last day grandfathered Pro 200 subscribers keep their old included allowance. The price stays at $200; the usage you get for it roughly halves. (Source: OpenAI Help Center, About ChatGPT Pro tiers)

OpenAI announced the change at DevDay on September 29, alongside a new Pro 500 plan at $500 a month. With the cutoff 12 days away, here is what actually changes and how to pick a tier.

Key facts:

• Three Pro plans: Pro 100 ($100), Pro 200 ($200) and Pro 500 ($500), all billed monthly with no annual option.
• Pro 200 after the cutoff: included usage in ChatGPT Work and Codex falls from 20x to 10x the Plus allowance, and GPT-6 Pro messages in chat fall from 200 to 100 a week (The Next Web, Engadget).
• Who is grandfathered: anyone with an active Pro 200 subscription at any point from September 22, 2026 through 10 a.m. Pacific on September 29, 2026. They keep the old allowance through October 29 while the subscription stays active.
• Ultrafast is Pro 500 only: buying credits on Pro 100 or Pro 200 does not unlock it (ChatGPT release notes, September 29, 2026).

What Ultrafast really costs

OpenAI says Ultrafast generates up to 300 tokens per second in Codex, which The Next Web reports as up to eight times faster than Standard. The catch is in the allowance math on OpenAI’s help page: for GPT-6 Astra and GPT-6.1 Sol, Ultrafast consumes your included usage at 8x the Standard rate, then bills purchased credits at 6x the Standard credit rate once the allowance is gone. For comparison, Fast mode consumes allowance at 2.5x.

OpenAI also notes these are “allowance consumption rates, not guarantees of faster performance.” So Pro 500’s bigger allowance does not translate into more work if you leave Ultrafast on all day. It buys wall-clock time, and you pay for it in allowance.

Grandfathering details that trip people up

• Keeping the old allowance does not give you Pro 500 features. You are still on Pro 200, with no Ultrafast.
• If you cancel and resubscribe on the same account, you still qualify, but only if you had an active Pro 200 subscription in that September 22 to 29 window.
• Scheduling a cancellation does not end the plan immediately; you keep access until the end of the billing period and can undo it.
• According to The Next Web, existing subscribers also get a one-time grant of usage credits worth $2,500 that expires at the end of the year. Engadget describes it only as a one-time credit, so check the email OpenAI sends affected users for your exact amount.
• OpenAI told The Next Web it will not bring back Pro 200’s five-hour limit.

How to decide

OpenAI’s help page does not publish a usage multiplier for Pro 100 or Pro 500. It only says Pro 200 includes more than Pro 100, and Pro 500 has the most. So the practical test is your own history: before October 29, think about how often you hit your Codex or Work limit on the current 20x allowance in a typical week. (Credit balances and purchases live in Settings → Usage in ChatGPT, or Usage & Billing in Codex.)

• You rarely get near the limit: the new 10x Pro 200 may still cover you. Compare it with Pro 100 on the pricing page before renewing.
• Regularly near the cap: expect to buy credits on Pro 200 after the cutoff. Credits cover Codex, ChatGPT Work and the Office add-ins, and included usage is always spent first.
• You are blocked on agent latency, not volume: that is the only case where Pro 500 clearly makes sense, because Ultrafast is the one thing no other tier or credit purchase unlocks.

If you mostly call OpenAI models from your own code, the API is billed separately per token. See our GPT-6 Sol and Luna pricing breakdown and the GPT-6 Astra availability notes. For another vendor that moved coding tools to metered usage, see GitHub Copilot’s switch to AI Credits.