OpenAI has kept ChatGPT Pro 200 at $200 a month while reducing its included usage for new subscribers, and introduced Pro 500 at $500 a month with a larger allowance and exclusive access to Astra Ultrafast. Eligible existing Pro 200 subscribers have until October 30, 2026, before their allowance changes, but the transition does not permanently preserve the old limits.
1
2
3
What each Pro tier includes
| Plan |
Monthly price |
Included usage and features |
| Pro 200 |
$200 |
10× the ChatGPT Plus allowance for Work and Codex; up to 100 GPT-6 Pro chat messages per week; no Astra Ultrafast. 1 10 15 |
| Pro 500 |
$500 |
25× the Plus allowance and access to Astra Ultrafast. 1 3 |
OpenAI’s published limits for Pro 200 are half the former 20× Plus allowance for Work and Codex and half the former weekly GPT-6 Pro chat limit of 200 messages. The subscription price remains $200 a month.
1
10
15
When the lower Pro 200 limits apply
New Pro 200 subscriptions reopened with the reduced allowance on September 29, 2026. Existing subscribers who had an active Pro 200 subscription at any point from September 22 through 10 a.m. Pacific Time on September 29 can keep their previous allowance through October 29, as long as their subscription remains active. The lower allowance applies to them starting October 30.
1
2
3
Eligible subscribers are also receiving a one-time grant of 62,500 usage credits, described in reporting as worth $2,500. The grant is temporary—not a permanent extension of the old allowance—and reporting says the credits expire on December 31, 2026.
6
15
16
What Astra Ultrafast and the five-hour limit mean
Astra Ultrafast is a lower-latency mode for GPT-6 Astra, not a separate Pro plan. OpenAI lists it as included with Pro 500; Pro 200 does not include access to it.
1
7
Pro plans currently have no five-hour usage limit in Work and Codex. That does not mean usage is unlimited: each plan still has an included allowance, and the plan’s allowance remains the constraint.
7
What SemiAnalysis’s comparison says about value
SemiAnalysis tested subscription limits using demanding coding and agent-style tasks and compared the resulting usage with what similar usage would cost at API prices. Its reported comparison found that Anthropic’s $200 Claude subscription offered roughly five times the API-equivalent value of OpenAI’s $200 tier on the models and workloads tested.
18
19
That figure is a comparison of usage valued at API rates. It is not a score of which model is better, nor a calculation of what either company actually spends to serve a subscriber.
18
30
The comparison helps explain why usage allowances matter as much as the monthly price for people who regularly run intensive tasks. For Pro 200 subscribers, the practical change is straightforward: the $200 price stays the same, but the included allowance falls; the credit grant and transition window soften that change without reversing it.
1
16
19