Codex Usage Limits: Pro, Plus & Business Compared
Table of Contents
- ChatGPT Plus Codex Quotas
- ChatGPT Business Codex Quotas
- ChatGPT Pro Codex Quotas
- Pro, Plus & Business Codex 5-Hour Quotas Compared
- How to Check Your Codex Quota
- Practical Tips to Reduce Codex Consumption
Check the latest Codex usage limits: want to quickly get the up-to-date 5-hour Codex quotas for ChatGPT Pro, Plus and Business? Based on official OpenAI information released on April 2, 2026, this article compares the Codex quota changes across the three memberships in detail and walks you through checking your own 5-hour Codex quota for ChatGPT Pro, Plus and Business, step by step. Whether you're a heavy coding user or a day-to-day developer, you'll find the membership that fits you best!
Starting April 2, OpenAI began cutting the price of Business from $30/month to $25/month. At first glance that looks like good news, but anyone using Codex recently has probably noticed that Business quota runs out fast.
Yes — from my investigation, OpenAI also quietly reduced Business's 5-hour Codex quota and weekly quota.
ChatGPT Plus Codex Quotas
According to the official docs, the 5-hour limits for local messages on the Codex CLI we commonly use are:
- GPT-5.4 quota: 33 to 168 messages
- GPT-5.4-mini quota: 110 to 560 messages
- GPT-5.3-Codex quota: 45 to 225 messages
ChatGPT Business Codex Quotas
Business memberships come with discounts — most people get a discounted Business plan for Codex, costing anywhere from tens to a hundred-plus yuan.
The most stable option in daily use is the 99-yuan Business membership. The currently published 5-hour Codex quotas are:
- GPT-5.4 quota: 15 to 60 messages
- GPT-5.4-mini quota: 40 to 200 messages
- GPT-5.3-Codex quota: 20 to 90 messages
Benefit #1 of Business: it comes with 15 extra uses of the Pro model. Previously, the Pro model was only available to the $200 Pro membership, but Business gets to try it — and 15 uses is nothing to sneeze at. A single Pro run consumes a very expensive amount of tokens, and one Pro task normally takes over an hour; I typically wait 70 to 90 minutes for a result, but the quality is generally higher.
Benefit #2 is the cheaper price. If you also need to upgrade ChatGPT Plus or Business at a discounted rate, you can use this system (stable, legitimate channels are rare, so we recommend bookmarking it):
Properly upgrading Plus never leads to account bans. If you've been banned, it's almost always because you ran into a shady "recharge" service or a zero-cost scam. This system has been running stably for a long time, and some renewing users have stayed for over a year.
GPT one-click upgrade system: gptplus.org.cn
ChatGPT Pro Codex Quotas
For heavy Codex users who want more convenience, the answer is the Pro membership — though the price is steep.
The 5-hour Codex quotas for Pro, as currently published by OpenAI, are:
- GPT-5.4 quota: 223 to 1120 messages
- GPT-5.4-mini quota: 743 to 3733 messages
- GPT-5.3-Codex quota: 300 to 1500 messages
Pro, Plus & Business Codex 5-Hour Quotas Compared
Below I've calculated the 5-hour Codex quota multipliers between Pro, Plus and Business, using Plus as the baseline:
| Model | Plus (baseline) | Business multiplier | Pro multiplier |
|---|---|---|---|
| GPT-5.4 | 1x | 0.36-0.45x | ≈ 6.7x |
| GPT-5.4-mini | 1x | ≈ 0.36x | ≈ 6.7x |
| GPT-5.3-Codex | 1x | 0.40-0.44x | ≈ 6.67x |
So Business gets roughly 60% cut, while Pro gets a bump — instead of about 5x the Plus quota before, it's now 6.7x the Plus quota.
Pro currently costs 10x as much as Plus, but factoring in the chat-side models, Pro is still worth it — you get unlimited use of the Pro model and high-quota deep research, among other features.
Based on the data, Plus is currently the membership with the best Codex quota value for money, so Plus is a solid first choice.
How to Check Your Codex Quota
On the web, you can check it via the Codex usage dashboard.
Path: https://chatgpt.com/codex/settings/usage
In the local Codex CLI, just run the /status command to check.
In the editor, use the local button in the bottom-left corner — click it to see the corresponding quota.
On the Codex App, it's in the same local button location. OpenAI is promoting the Codex app — if you use Codex in the app, your quota is doubled.
Practical Tips to Reduce Codex Consumption
If Codex feels slow, you can turn on Fast mode — it's much faster, but consumes double the quota.
Of course, you can also reduce 5-hour Codex quota consumption in these ways:
- Control your prompts. When asking Codex, be clear and complete, and start fresh windows as needed to drop unnecessary context.
- Shrink your AGENTS.md file. On larger projects, you can control how much context is injected by nesting AGENTS.md files within your repository.
- Limit the number of MCP servers you use. Every MCP you add to Codex adds more context to your messages and eats more of your quota. Disable MCP servers when you don't need them.
- Switch routine tasks to GPT-5.4-mini. Using a smaller model should extend your local message quota by roughly 2.5x to 3.3x, depending on which model you're switching from. (Per OpenAI.)
That said, when you're coding, use the best, most expensive model you can afford. Spending half a day fixing a bug also burns time and tokens — better to just get it right the first time.
Summary: with this comparison and checking guide for the 5-hour Codex quotas of ChatGPT Pro, Plus and Business, you should now know how to pick the most cost-effective membership. We recommend bookmarking this article and coming back to check the latest quota changes!