The 20-second mental model (so you stop getting confused)
OpenAI has multiple billing “buckets”:
- ChatGPT subscription (Free/Go/Plus/Pro/Business) → billed as a monthly subscription in ChatGPT.
- Codex “in ChatGPT” usage → has time/weekly limits and optional credits (a separate “flexible usage” concept). (OpenAI Help Center)
- Sora “in ChatGPT” usage → uses credits (also “flexible usage”). (OpenAI Help Center)
- API Platform usage (platform.openai.com) → pay-as-you-go by tokens/requests, tracked in the API Usage Dashboard. (OpenAI Help Center)
So ChatGPT subscription ≠ API credits. And Codex credits ≠ API usage.
Master FAQ (50 Qs + detailed answers)
A) Accounts, billing separation, and “where do I look?”
1) Are ChatGPT, Codex, Sora, and API billed together or separately?
They are tracked and billed separately depending on where you use them. ChatGPT plans are subscriptions, while API calls are metered usage, and Codex/Sora can use “credits for flexible usage.” That’s why you can see $0 on API usage but still hit Codex credits limits or API quota/billing errors. (OpenAI Help Center)
2) What’s the difference between an OpenAI “organization” and a “project”?
On the API Platform, usage, keys, budgets, and resources are organized into Organizations → Projects. API keys can be project-scoped, and roles/visibility can restrict who can see usage. (OpenAI Help Center)
3) Where do I check ChatGPT plan billing?
In ChatGPT: Settings → Account (and the plan management area). That shows your subscription status (Plus/Pro/etc.) and renewal.
4) Where do I check Codex usage and limits?
Codex has its own usage dashboard: chatgpt.com/codex/settings/usage (you already opened this). It shows the 5-hour limit, weekly limit, and credits remaining.
5) Where do I check Sora usage/credits?
In Sora (sora UI) and in the “credits / flexible usage” area associated with Sora. Sora has its own billing/credits rules. (OpenAI Help Center)
6) Where do I check API usage/spend?
On the API Platform Usage Dashboard: platform.openai.com/usage. It also has project filtering and uses UTC. (OpenAI Help Center)
7) Why can API usage look like $0 even though I’m getting “Quota exceeded”?
Because that error often happens before any billable usage can occur (for example: no billing set up / no prepaid credits / account not fully onboarded / wrong org). So there’s nothing to show as spend yet.
8) Can one dashboard show everything combined (ChatGPT + Codex + Sora + API)?
No single dashboard cleanly consolidates all of it. Even within the API platform, usage is not automatically consolidated across separate org structures. (OpenAI Help Center)
9) What are “credits” in ChatGPT/Codex/Sora?
They’re a “flexible usage” mechanism to continue using certain features beyond included plan limits. Credits are separate from API billing and follow their own terms (expiry, eligibility, etc.). (OpenAI Help Center)
10) Do credits expire?
Yes—OpenAI’s credit terms describe expiration/validity rules (and credits are not the same as “subscription time”). (OpenAI Help Center)
B) ChatGPT consumer & business plans (what you pay for)
11) What are the main ChatGPT plan types?
Typically: Free, Go, Plus, Pro for individuals, and Business, Enterprise for teams/orgs. Business/Enterprise feature matrices are shown on OpenAI’s business pricing pages. (OpenAI)
12) How is ChatGPT Free priced?
$0 subscription. You still get access but with tighter feature limits and availability compared with paid tiers.
13) What is ChatGPT Go and how much does it cost?
ChatGPT Go is a lower-cost paid tier; OpenAI’s announcement lists it at $8/month (region/taxes may vary). (OpenAI)
14) What is ChatGPT Plus and how much does it cost?
Plus is listed as $20/month (region/taxes may vary). (OpenAI Developers)
15) What is ChatGPT Pro and how much does it cost?
Pro is listed as $200/month (region/taxes may vary). (OpenAI Developers)
16) What is ChatGPT Business pricing?
Business is priced per user (and varies by billing cadence/region). For example, OpenAI’s Business pricing page shows Business and Enterprise feature comparisons and purchase links; local currency pricing can appear (you may see JPY in Japan). (OpenAI)
17) What’s ChatGPT Enterprise pricing?
Enterprise is typically sales-led / contract pricing (“Contact sales”). (OpenAI)
18) What does “Unlimited” vs “Flexible” mean on Business/Enterprise feature tables?
“Unlimited” generally means included usage for that feature; “Flexible” indicates usage that can expand with policies/credits/contract terms rather than a simple unlimited bucket. OpenAI uses these labels in the plan comparison tables. (OpenAI)
19) Do ChatGPT subscriptions include API usage?
No—ChatGPT subscription is separate from API metered billing. API usage is tracked in the API Usage Dashboard and requires API billing setup. (OpenAI Help Center)
20) If I pay for ChatGPT (Plus/Pro), do I automatically get Codex/Sora credits?
You get whatever your plan includes in terms of included limits and eligibility for “flexible usage.” The authoritative place to see your included balances is the Codex/Sora usage pages and the “credits for flexible usage” documentation. (OpenAI Help Center)
C) Codex pricing & limits (and why it says “out of credits”)
21) What is Codex in 2026 (in OpenAI terms)?
Codex is an AI coding agent experience (web/IDE/CLI) that can run tasks like code edits, diffs, tests, and code review.
22) Is Codex billed like API tokens or like ChatGPT messages?
It depends how you use it:
- In the Codex product experience, you hit time/weekly limits and may use credits for flexible usage. (OpenAI Help Center)
- If you use Codex models via the API, that’s API token billing (metered on the API platform). (OpenAI Developers)
23) What is the “5-hour usage limit” in Codex?
It’s a rolling allowance used by Codex agent sessions (you saw it in your Codex usage dashboard). When you exceed it, you may need to wait for reset or use credits (if available).
24) What is the “weekly usage limit” in Codex?
Same idea, but measured on a weekly window to prevent unlimited heavy usage in a short period.
25) What are “Codex credits”?
Credits let you keep sending Codex messages / tasks beyond plan limits. Your Codex dashboard shows Credits remaining (you had 0). (OpenAI Help Center)
26) “ChatGPT gave some message credit to Codex—how much?”
It’s not a fixed universal number across everyone. The exact balance depends on your plan, your usage, and any credit rules/promos at that time. The only reliable answer is your Codex usage dashboard, which shows remaining credits and limits. (OpenAI Help Center)
27) Why does Codex say “out of credits” but API usage is $0?
Because Codex credits are not API billing. Codex flexible usage credits are a separate bucket from API tokens/spend. So API can still show $0 even if Codex credits are depleted. (OpenAI Help Center)
28) Can I “fix” Codex out-of-credits by adding money to API billing?
Not necessarily. Adding API billing helps with API calls, not automatically Codex flexible credits (they are managed under the credits/flexible usage system). (OpenAI Help Center)
29) What does this Codex error mean: “Encrypted content … organization_id did not match”?
That message usually indicates you authenticated or stored encrypted session data under one org/account, then tried to use it under another org/account target—so the encrypted data can’t be validated for the new org. (This exact error has been reported in Codex tooling contexts.) (GitHub)
Typical fix: sign out, clear local Codex auth/session cache, and re-authenticate using the same org/account you intend to bill against.
30) Why does Codex CLI say “Quota exceeded. Check your plan and billing details”?
If the CLI is making API calls, that error often means the API org/project has no billing/credits set, or you’re using a key tied to a different org/project than you think. The API Usage Dashboard and project/org settings are where you validate this. (OpenAI Help Center)
D) Sora pricing & credits (consumer) + Sora in Business plans
31) Is Sora billed by subscription, credits, or API usage?
All of the above exist in different contexts:
- Sora consumer app typically uses credits / plan entitlements. (OpenAI Help Center)
- Sora via API is metered on the API pricing pages for video/image related models (API billing). (OpenAI Developers)
32) Does ChatGPT Business include Sora?
OpenAI’s Business feature comparison explicitly lists “Sora 1: Limited” for Business and not included for Enterprise (in that table). (OpenAI)
33) How do Sora credits work?
You consume credits when generating videos; credit burn depends on factors like duration/quality. The official Sora billing FAQ + credits documentation describes how credits apply. (OpenAI Help Center)
34) Do Sora credits expire?
Credits follow OpenAI credit terms (validity/expiration rules apply). (OpenAI Help Center)
35) Can I buy more Sora credits any time?
Depending on your plan/region/eligibility, you can often add credits for flexible usage; the credits system is described in OpenAI’s flexible usage documentation. (OpenAI Help Center)
36) Why might Sora not appear on my Enterprise/Edu account?
Sora availability can differ by plan and eligibility; OpenAI’s Sora billing FAQ notes plan-based access constraints. (OpenAI Help Center)
E) API Platform pricing (tokens), limits, budgets, and why you see $0
37) How does API pricing work at the core level?
API is billed mainly by tokens: you pay for input tokens, sometimes cached input, and output tokens per model. The official API pricing table lists exact rates. (OpenAI Developers)
38) What are current example token prices for GPT-5.2?
From OpenAI’s pricing table (per 1M tokens):
- Flex: gpt-5.2 input $0.875, cached input $0.0875, output $7.00 (OpenAI Developers)
- Standard: gpt-5.2 input $1.75, cached input $0.175, output $14.00 (OpenAI Developers)
(There’s also a higher tier section with higher prices, typically used for higher-priority processing. (OpenAI Developers))
39) What are Flex vs Standard (and the higher-priced tier) in API pricing?
OpenAI publishes separate sections: Flex (lower cost), Standard (normal), and an additional higher-priced section (used when you want higher priority / faster scheduling). The tradeoff is generally cost vs performance/priority. (OpenAI Developers)
40) What is “cached input” and why is it cheaper?
Cached input is discounted input pricing used when repeated prompt fragments are served from cache (same content reused). The pricing tables show cached input rates lower than standard input. (OpenAI Developers)
41) How is Codex billed on the API (model names like gpt-5.2-codex)?
The API pricing table lists codex model variants (e.g., gpt-5.2-codex) with the same input/cached/output structure under pricing tiers. (OpenAI Developers)
42) How is image generation billed on the API now?
OpenAI lists image tokens pricing (per 1M tokens) for image models (e.g., gpt-image-1, chatgpt-image-latest) with Batch/Standard rates. (OpenAI Developers)
43) Give an example of image token prices.
For image tokens (per 1M tokens), OpenAI lists (examples):
- Standard: gpt-image-1 input $10.00, cached input $2.50, output $40.00 (OpenAI Developers)
- Batch (discounted): gpt-image-1 input $5.00, cached input $1.25, output $20.00 (OpenAI Developers)
44) How is Realtime API billed?
Realtime models are listed with their own pricing lines (input/cached input shown). (OpenAI Developers)
45) Why does the API Usage Dashboard show UTC and feel “off by a day”?
Because usage dashboard data is displayed using UTC, not your local timezone. (OpenAI Help Center)
46) Why are there two project selectors on the API usage page?
OpenAI explains that the usage page has its own selector independent from the top bar project selection for more flexible filtering. (OpenAI Help Center)
47) Where do I see “credit grants” (API credits / prepaid credits) on the platform?
OpenAI says the credit grants table is unified under the organization billing page, and points to the billing/credit-grants section. (OpenAI Help Center)
48) How do API budgets work—do they hard-stop requests?
Budgets are described as soft thresholds for monitoring/alerts; exceeding them can still allow requests to process (it’s for monitoring rather than a guaranteed hard stop). (OpenAI Help Center)
49) What causes “Quota exceeded. Check your plan and billing details” on the API?
Common causes:
- No billing method / no prepaid credits enabled for that org/project
- Using an API key tied to a different org/project than the one you’re checking
- Actual spend/credit exhaustion in that billing bucket
You diagnose it by checking org/project, billing, and usage visibility settings. (OpenAI Help Center)
50) Why do org/project mismatches happen so easily (especially with multiple accounts)?
Because keys, projects, and visibility are scoped. OpenAI also notes the usage dashboard doesn’t consolidate across org structures automatically, and project scoping + role/visibility settings can hide usage from some users. (OpenAI Help Center)
Your specific situation (based on what you shared)
- Your API Platform screenshot shows $0 usage/spend.
- Your Codex usage dashboard shows Credits remaining = 0 (and some usage bars).
- Your API call error “Quota exceeded…” is consistent with API billing not enabled / wrong org/project / no API credits, which can happen even if “usage” is currently $0 (because requests are blocked early). (OpenAI Help Center)
- Your org mismatch encrypted content error is also consistent with signing into Codex/CLI under one org/account and then trying another. (GitHub)
One important note: you pasted a live API key in chat. You should rotate/delete that key immediately in the API Platform and create a new one (treat it as compromised).