OpenAI Quota Query: 3-Hour & 24-Hour Limits

What You'll Learn

• Use /mystatus to query OpenAI Plus/Team/Pro subscription quotas
• Understand 3-hour and 24-hour quota limit information in the output
• Distinguish between primary and secondary windows
• Handle token expiration properly

Your Current Challenge

OpenAI API calls have rate limits, and exceeding them triggers temporary access restrictions. But you don't know:

• How much quota remains?
• Which window is active—the 3-hour or 24-hour one?
• When will it reset?
• Why do you sometimes see data for two windows?

Without timely access to this information, your progress on coding with ChatGPT or working on projects may be impacted.

When to Use This

When you:

• Need to frequently use the OpenAI API for development
• Notice slower responses or rate limiting
• Want to understand team account usage
• Need to know when quotas will refresh

Core Concepts

OpenAI applies two types of rate limiting windows for API calls:

Window Type	Duration	Purpose
Primary Window	Returned by OpenAI server	Prevents excessive calls within a short time
Secondary Window	Returned by OpenAI server (may not exist)	Prevents long-term overuse

mystatus queries both windows in parallel and displays:

• Percentage used
• Remaining quota progress bar
• Time until reset

INFO

Window duration is returned by the OpenAI server and may vary by subscription type (Plus, Team, Pro).

Follow Along

Step 1: Execute the Query Command

In OpenCode, type /mystatus. The system automatically queries quotas for all configured platforms.

You should see: Quota information for OpenAI, Zhipu AI, Z.ai, Copilot, Google Cloud, and other platforms (depending on which platforms you've configured).

Step 2: Locate the OpenAI Section

Find the ## OpenAI Account Quota section in the output.

You should see: Something like this:

## OpenAI Account Quota

Account:        [email protected] (plus)

3-hour limit
███████████████░░░░░░░░░░ 60% remaining
Resets in: 2h 30m

Step 3: Interpret the Primary Window

The Primary Window (primary_window) typically shows:

• Window name: Such as 3-hour limit or 24-hour limit
• Progress bar: Visual representation of remaining quota ratio
• Remaining percentage: Such as 60% remaining
• Reset time: Such as Resets in: 2h 30m

You should see:

• Window name displays the duration (3-hour / 24-hour)
• A fuller progress bar means more quota remaining; an emptier bar means closer to depletion
• Reset time is a countdown—quota refreshes when it reaches zero

WARNING

If you see Limit reached!, the current window quota is exhausted and you need to wait for reset.

Step 4: Check the Secondary Window (if present)

If OpenAI returns secondary window data, you'll see:

24-hour limit
████████████████████████████ 90% remaining
Resets in: 20h 45m

You should see:

• Secondary window shows quota for another time dimension (usually 24 hours)
• May have a different remaining percentage than the primary window

TIP

The secondary window is an independent quota pool. Exhausting the primary window doesn't affect the secondary window, and vice versa.

Step 5: View Subscription Type

You can see the subscription type in the Account line:

Account:        [email protected] (plus)
                                   ^^^^^
                                   Subscription type

Common subscription types:

• plus: Personal Plus subscription
• team: Team/organization subscription
• pro: Pro subscription

You should see:

• Your account type displayed in parentheses after the email
• Different subscription types may have different limits

Checkpoint ✅

Verify your understanding:

Scenario	You should see
60% remaining in primary window	Progress bar about 60% full, displaying 60% remaining
Resets in 2.5 hours	Shows Resets in: 2h 30m
Limit reached	Displays Limit reached!
Secondary window present	Both primary and secondary windows show their own data lines

Common Pitfalls

❌ Mistake: Not Refreshing After Token Expiration

Error symptom: You see the message ⚠️ OAuth token expired (English) or ⚠️ OAuth 授权已过期 (Chinese)

Cause: The OAuth Token has expired (specific duration controlled by the server), preventing quota queries.

Correct approach:

1. Re-login to OpenAI in OpenCode
2. Token refreshes automatically
3. Run /mystatus query again

❌ Mistake: Confusing Primary and Secondary Windows

Error symptom: Thinking there's only one window quota, then the primary window is exhausted while the secondary window is still available

Cause: The two windows are independent quota pools.

Correct approach:

• Pay attention to reset times for both windows
• Primary window resets faster; secondary window resets slower
• Distribute usage reasonably to avoid long-term overage in any window

❌ Mistake: Ignoring Team Account ID

Error symptom: Team subscription displays usage that isn't yours

Cause: Team subscriptions require passing a team account ID; otherwise, the query might be for the default account.

Correct approach:

• Ensure you're logged into the correct team account in OpenCode
• Token automatically includes chatgpt_account_id

Lesson Summary

mystatus queries quotas by calling the official OpenAI API:

• Supports OAuth authentication (Plus/Team/Pro)
• Displays primary and secondary windows (if present)
• Progress bars visualize remaining quota
• Countdown shows reset time
• Automatically detects token expiration

Coming Up Next

In the next lesson, we'll learn about Zhipu AI and Z.ai Quota Query.

You'll learn:

• What the 5-hour Token limit means
• How to check MCP monthly quotas
• Warning prompts when usage exceeds 80%

---

Appendix: Source Code Reference

Click to expand source code locations

Last updated: 2026-01-23

Function	File Path	Lines
OpenAI quota query entry	plugin/lib/openai.ts	207-236
OpenAI API call	plugin/lib/openai.ts	132-155
Format output	plugin/lib/openai.ts	164-194
JWT Token parsing	plugin/lib/openai.ts	64-73
Extract user email	plugin/lib/openai.ts	78-81
Token expiration check	plugin/lib/openai.ts	216-221
OpenAIAuthData type definition	plugin/lib/types.ts	28-33

Constants:

• OPENAI_USAGE_URL = "https://chatgpt.com/backend-api/wham/usage": Official OpenAI quota query API

Key functions:

• queryOpenAIUsage(authData): Main function for querying OpenAI quota
• fetchOpenAIUsage(accessToken): Call the OpenAI API
• formatOpenAIUsage(data, email): Format output
• parseJwt(token): Parse JWT Token (non-standard library implementation)
• getEmailFromJwt(token): Extract user email from Token
• getAccountIdFromJwt(token): Extract team account ID from Token