Codex Cloud pricing: what cloud tasks cost, the free tier, and how usage is counted
On September 29, 2026, Codex went properly cloud-native: you can hand a task to Codex from your computer, your phone, or a browser, and the task keeps working in a cloud environment while your laptop is asleep — reusable environments, shared team setups, review and merge from anywhere (OpenAI: DevDay 2026 Recap (Sep 29, 2026) OpenAI Learn: Codex Cloud docs). The pricing question that follows is simpler than it looks: Codex Cloud is not sold separately. Codex is included in ChatGPT plans, and cloud tasks spend the same usage allowance as everything else you run in Codex (OpenAI Learn: Codex pricing (plans, free tier, limits)). What changes in the cloud is not the price list — it is how fast the allowance drains.
What Codex Cloud costs on each plan
OpenAI’s Codex pricing page is explicit: “ChatGPT Work and Codex share usage” — the same pricing, credits and usage limits apply, and API token prices are separate from subscription usage (OpenAI Learn: Codex pricing (plans, free tier, limits)). There is no Codex Cloud SKU:
| Plan | Price | What you get for Codex |
|---|---|---|
| Free | $0/month | GPT-6 Luna at Standard speed in the desktop appQuick coding tasks, subject to rollout |
| Go | $8/month | GPT-6 Luna at Standard speed in the desktop appLightweight coding tasks |
| Plus | $20/month | Codex on the web, CLI, IDE extension and iOSGPT-6 Sol and GPT-6 LunaCloud-based integrations (automatic code review, Slack)Extend usage with ChatGPT credits |
| Pro | $100 / $200 / $500 per month | Everything in PlusUltrafast on Pro 500No five-hour limit — the allowance is the constraint |
| Business | $20/user/month (annual; $25 monthly) | Codex across desktop and mobileLarger cloud virtual machinesExtend with credits |
| Enterprise & Edu | Contact sales | Priority request processingCredit-based flexible pricing or per-seat limits |
| API key | Pay per API price list | Codex in CLI, SDK and IDE extensionNo cloud features (GitHub code review, Slack, etc.)Billed at API token rates, not plan usage |
OpenAI Learn: Codex pricing (plans, free tier, limits) OpenAI Help Center: About ChatGPT Pro tiers. Two rows deserve a second look. The API key row is the only true pay-as-you-go path: it follows API token prices and gets no cloud features. And Free really does include Codex — but only as GPT-6 Luna at Standard speed in the desktop app, aimed at quick coding tasks and subject to rollout (OpenAI Learn: Codex pricing (plans, free tier, limits)).
How cloud tasks eat your allowance
The metering rules OpenAI publishes, verbatim in spirit:
- One shared allowance. “Local messages and cloud chats share your plan’s usage allowance. Weekly limits may also apply.” (OpenAI Learn: Codex pricing (plans, free tier, limits))
- Cloud can cost more per task. “Cloud tasks may use more of your allowance than local messages.” (OpenAI Learn: Codex pricing (plans, free tier, limits))
- Task shape beats prompt length. Model choice, context size, reasoning, tool use, retrieval and caching all affect usage — “prompt length alone isn’t a reliable estimate” (OpenAI Learn: Codex pricing (plans, free tier, limits)).
- Speed modes multiply the burn. Fast mode spends included usage at 2.5× the standard rate; GPT-6 Astra Ultrafast at 8× (6× for purchased credits). The multipliers describe usage burn, not speed. (OpenAI Learn: Codex pricing (plans, free tier, limits) OpenAI Learn: ChatGPT usage limits and spend controls)
- Pro has no five-hour wall. The published message estimates are per five-hour period for Plus and Standard Business; “Pro plans currently have no five-hour limit.” (OpenAI Learn: Codex pricing (plans, free tier, limits) OpenAI Help Center: About ChatGPT Pro tiers)
For calibration, OpenAI’s published estimates for local messages per five-hour period on Plus and Standard Business (OpenAI Learn: Codex pricing (plans, free tier, limits)):
| Model | Plus (local msgs / 5h) | Standard Business (local msgs / 5h) |
|---|---|---|
| GPT-6 Astra | 5–45 | 5–45 |
| GPT-6.1 Sol | 15–160 | 15–160 |
| GPT-6 Sol | 15–150 | 15–150 |
| GPT-6 Luna | 350–3,000 | 350–3,000 |
These are estimates, not fixed limits — “cloud tasks may use more” applies on top. If a long cloud task must hold a large context for hours, expect it to weigh far more than a chat-sized exchange.
Is there a Codex Cloud free tier?
Yes, with hard edges. The ChatGPT Free plan ($0) includes Codex on the desktop app running GPT-6 Luna at Standard speed, framed for “quick coding tasks” and still subject to rollout (OpenAI Learn: Codex pricing (plans, free tier, limits)). The Free plan has no cloud environment story at that tier — no web Codex, no long-running cloud tasks, and image generation isn’t on Free either (OpenAI Learn: Codex pricing (plans, free tier, limits)). DevDay’s cloud announcement lists availability from Plus upward: Plus, Pro, Business, Healthcare, Education and Enterprise (OpenAI: DevDay 2026 Recap (Sep 29, 2026)). So: the free tier is a taste, the cloud starts at $20.
What happens when you hit the limit
OpenAI wants in-flight work to finish: if you reach your limits during an active turn, the agent can continue that turn, subject to fair use. After that, Plus and Pro users can buy credits to keep going without upgrading; Business, Edu and Enterprise workspaces on flexible pricing can buy workspace credits; and anyone can run extra local chats with an API key, billed at standard API rates (OpenAI Learn: Codex pricing (plans, free tier, limits)). Limits and reset times live in the usage dashboard at chatgpt.com/codex/settings/usage, or /status inside a Codex CLI session (OpenAI Learn: Codex pricing (plans, free tier, limits)).
Track the reset clocks behind cloud tasks
Because cloud tasks can drain an allowance asynchronously — a task you started from your phone at lunch may be the reason Codex is quiet at dinner — the reset time is the number that matters. ChatGPT shows it when available, and Support cannot reset limits for you (OpenAI Help Center: About ChatGPT Pro tiers). QuotaClock turns it into a countdown you can leave open:
- Read the reset time in the usage dashboard or in ChatGPT (Settings → Usage).
- Register it in the countdown tool: set Window started at to now, type the minutes until reset as the window length, and name it “Codex cloud”.
- Confirm at zero in the dashboard — a zero proves your arithmetic, not OpenAI’s reset.
Related pages: Codex Security Cloud pricing and quota use (the other big new consumer), the ChatGPT Pro 500 vs 200 vs 100 comparison if you are deciding which ceiling to buy, and the ChatGPT dots limits page for what dots spend. How QuotaClock works explains why the tool never touches your account.
Frequently asked questions
Is Codex Cloud free?
There is no separate Codex Cloud price — cloud tasks are included in ChatGPT plans from Plus upward, and they spend your plan’s shared Codex usage. The Free plan gets desktop Codex with GPT-6 Luna for quick tasks only; beyond included usage you wait for resets or buy credits.
Does a cloud task use more usage than a local one?
It can. OpenAI states that local messages and cloud chats share one allowance and that cloud tasks may use more of it than local messages. Context size, model, tool use and validation work all move the number.
What is the Codex free tier exactly?
ChatGPT Free includes Codex on the desktop app: GPT-6 Luna at Standard speed, aimed at quick coding tasks, with rollout still in progress. Web Codex, cloud environments and integrations start at Plus ($20/month).
Where can I see my Codex Cloud usage and limits?
In the usage dashboard at chatgpt.com/codex/settings/usage — limits, remaining capacity and reset times — or with /status during a Codex CLI session. OpenAI suggests checking it every week or two to learn your pace.
Can I keep working after hitting my Codex limit?
The current turn can finish, subject to fair use. After that: wait for the reset, buy credits (Plus/Pro, and flexible workspaces), or run extra local chats with an API key billed at API rates.
Does Codex Cloud replace the Codex CLI or IDE extension?
No — they interoperate. You start tasks from the web, desktop or phone and review results there; the CLI and IDE extension remain the local surfaces, and an API key only covers those local surfaces, not cloud features.
Sources
All official OpenAI pages, last read 2026-09-30:
- OpenAI: DevDay 2026 Recap (Sep 29, 2026)
- OpenAI Learn: Codex Cloud docs
- OpenAI Learn: Codex pricing (plans, free tier, limits)
- OpenAI Learn: ChatGPT usage limits and spend controls
- OpenAI Help Center: About ChatGPT Pro tiers
QuotaClock is not affiliated with, endorsed by, or connected to OpenAI. “Codex” and “ChatGPT” are OpenAI product names, used here only to describe them.
QuotaClock is an independent, non-official tool. It is not affiliated with, endorsed by, or connected to Anthropic, OpenAI, or any other AI provider. It does not sign in to your accounts, does not read API keys, tokens, or usage data, cannot reset anything for you, and cannot promise that a provider's window rules stay the same. The countdown is built only from the times and lengths you type in. Window lengths and reset rules are set by each provider and change over time — always check the provider's own current documentation.