Is OpenAI Codex Free With ChatGPT Plus? Limits, Cloud Access, and API Costs

Is OpenAI Codex free with ChatGPT Plus? The short answer is yes: Codex access is included with ChatGPT plans, including Plus. However, the subscription does not mean unlimited usage, and it is important to separate local Codex work, Codex Cloud tasks, and API billing before choosing a workflow.

Access type	What it covers	Key limitation or cost
ChatGPT Plus	Codex on the web, CLI, IDE extension and iOS	$20 per month; usage varies by task and model
Codex Cloud	Remote coding tasks on OpenAI-managed computers	Available to eligible Plus accounts; rollout and workspace settings apply
API access	Developer-controlled requests using an API key	Billed separately according to token usage

What ChatGPT Plus actually includes

OpenAI’s Help Center documentation says Codex is included across ChatGPT plans, while Codex Cloud is available to eligible Plus, Pro, Business, Enterprise, Healthcare and Education accounts. Free and Go plans do not include Codex Cloud. For Plus users, cloud tasks can run on OpenAI-managed computers and continue while the user’s computer is asleep, subject to availability and workspace controls.

That makes Plus useful for developers who want to delegate repository tasks without keeping a local machine active. OpenAI’s cloud-environment documentation lists default Plus resources of two virtual CPUs, 8 GiB of memory and 8 GiB of disk. Saved virtual-machine state may be recoverable for up to seven days, while computer and browser use is currently unsupported in cloud environments.

Limits depend on how Codex is used

There is no single universal “Codex limit” that applies to every Plus subscriber. Usage can change with the selected model, task complexity, context size, reasoning effort, tools and whether work runs locally or in the cloud. Check the usage dashboard for the current allowance before planning a large project. Teams evaluating AI-assisted application work can also review project consultation services to map cloud tasks to a broader delivery workflow.

ChatGPT Plus is not the same as API access

A ChatGPT Plus subscription and an OpenAI API account are separate billing paths. Plus covers Codex access through supported ChatGPT experiences, but developers who authenticate with an API key pay according to API token usage. API billing does not automatically unlock Codex Cloud features, so an API budget should be treated as a separate engineering expense.

OpenAI’s current API pricing lists gpt-5.3-codex at $1.75 per 1 million input tokens, $14 per 1 million output tokens and $0.175 per 1 million cached input tokens. Those rates are not a flat monthly Codex fee. The final cost depends on the model selected, the amount of context sent, generated output and how much cached context can be reused. A short coding request and a long repository-analysis task can therefore have very different API costs.

How to plan a Plus-based Codex workflow

For individual development, start by identifying which work should remain local and which tasks benefit from a remote environment. Local work may be convenient for rapid edits, while cloud tasks are useful when a repository needs background analysis or implementation after the user leaves the computer. Before assigning a large task, confirm the current allowance in the usage dashboard and check whether the relevant workspace has enabled cloud access.

• Use Plus: when you want Codex within supported ChatGPT interfaces and eligible cloud-task access.
• Budget API usage separately: when an application, script or internal tool sends requests with an API key.
• Review the environment first: default Plus cloud resources are limited, and browser or computer interaction is currently unsupported there.

Businesses that need to connect AI-assisted coding with requirements, testing and deployment planning may benefit from a broader project consultation workflow. That helps separate subscription access from the operational decisions that determine project cost, review effort and release readiness.

What Plus users should check before starting

The most important question is not simply whether Codex is included, but which execution path a task will use. A local session, a ChatGPT-hosted cloud task and an API request can involve different allowances, environments and costs. Treat them as separate parts of the same development process rather than interchangeable access modes.

Match the task to the environment

Local Codex work is generally the better fit when the developer needs immediate access to local files, tools or an existing runtime. Cloud tasks are more suitable for repository work that can be delegated and reviewed later, provided the workspace permits them. Because OpenAI’s documented Plus cloud environment has limited compute, memory and disk, large builds or data-heavy workflows may need additional planning.

The current cloud documentation also says browser and computer use is unsupported in these environments. That matters for tasks that depend on visual interfaces, manual website interaction or desktop applications. In those cases, developers may need a local workflow or a separate testing process. Teams planning a broader implementation can review application development services when deciding how AI-assisted coding should fit alongside engineering review and deployment.

Keep subscription and project budgets separate

For a personal workflow, the Plus subscription is the predictable baseline expense, while usage limits remain variable. For an application that calls a model through an API key, estimate token consumption independently. Repository context, repeated instructions and long generated patches can increase usage, even when the individual requests appear simple.

• Check the usage dashboard before assigning a large cloud task.
• Confirm that the selected workspace and account are eligible for Codex Cloud.
• Review generated changes locally before merging or deploying them.
• Track API requests separately from ChatGPT subscription usage.

This separation makes it easier to forecast engineering effort, identify unexpected API spend and decide when a task requires human review or a more capable development environment.

What Codex access means for different users

The answer to “Is OpenAI Codex free with ChatGPT Plus?” depends on which Codex experience you mean. Plus provides access to supported Codex interfaces and, for eligible accounts, Codex Cloud. It does not create an unlimited pool of compute, remove usage limits or include separate API credits.

Workflow	What to expect	Budget consideration
Local Codex session	Works with the developer’s local files and tools	Covered by the relevant ChatGPT access, subject to limits
Codex Cloud	Runs tasks on OpenAI-managed computers and can continue while the computer is asleep	Eligible Plus access, rollout status and usage allowances apply
API-key request	Uses OpenAI models through a separate developer integration	Token-based API billing applies independently

When Codex Cloud is useful—and when it is not

Cloud execution is most useful when a task can be delegated, tested and reviewed later. For example, a developer might ask Codex to inspect a repository, propose a change or work through a defined implementation while other work continues locally. This can make background engineering tasks easier to organise, but the resulting changes still require review before merging.

OpenAI’s documented default Plus environment includes 2 virtual CPUs, 8 GiB of memory and 8 GiB of disk. Saved virtual-machine state may be recoverable for up to seven days. Those resources are suitable for bounded repository tasks, but they should not be treated as a general-purpose production server. Browser and computer interaction is also currently unsupported in cloud environments, limiting workflows that depend on visual websites or desktop software.

Teams combining AI-assisted coding with requirements, testing and release planning can use a structured project consultation process to define approval points, environment needs and ownership before work begins.

A practical cost-control checklist

• Separate the Plus subscription from API spending.
• Estimate context and output volume before using an API key.
• Reserve cloud tasks for work that benefits from background execution.
• Review generated code, tests and dependencies before deployment.

How API costs change the answer

For anyone asking, “Is OpenAI Codex free with ChatGPT Plus?”, the subscription answer and the API answer should remain separate. ChatGPT Plus is listed at $20 per month, but that fee does not include a separate balance for applications making API requests. API usage is billed according to the model and the number of tokens processed.

OpenAI’s current API pricing lists gpt-5.3-codex at $1.75 per million input tokens, $14 per million output tokens and $0.175 per million cached input tokens. These are usage rates, not a fixed project price. A repository task with extensive context, repeated prompts or large generated patches can consume more tokens than a short coding request. Teams should therefore estimate request volume and monitor actual usage before treating the published rates as a complete budget.

What Plus does—and does not—cover

• Included access: supported Codex experiences through eligible ChatGPT plans, subject to usage limits and account settings.
• Cloud execution: available only to eligible plans and workspaces, with rollout conditions and resource limits.
• API integration: authenticated and billed separately, with no automatic transfer of ChatGPT subscription benefits.
• Unlimited development capacity: not included; model, task complexity, context and execution path affect availability.

Before adopting Codex for a customer-facing product, define who approves generated changes, where secrets are stored and how tests are run. A documented project consultation service can help teams map those controls to requirements, environments and release responsibilities.

Questions to resolve before scaling usage

Check the ChatGPT usage dashboard for current allowances instead of relying on an old limit quoted elsewhere. Confirm that the relevant workspace has Codex Cloud enabled, then test a representative task with realistic repository context. Finally, compare the resulting review time and API spend with the value of delegating the work. This evidence-based approach gives Plus users a clearer basis for deciding which tasks belong locally, in the cloud or behind a separately funded API integration.

The bottom line: Is OpenAI Codex free with ChatGPT Plus?

Yes, Codex access is included with ChatGPT Plus, but “free” has important boundaries. The subscription provides supported Codex experiences and may provide Codex Cloud for eligible accounts. It does not mean unlimited usage, unlimited cloud capacity or free access to every development workflow.

For many developers, Plus is best understood as a convenient way to use Codex within ChatGPT’s access and allowance framework. Local work can remain close to existing files and tools, while cloud tasks can handle selected repository work in an OpenAI-managed environment. The right choice depends on task size, review requirements, repository complexity and whether the work needs browser or computer interaction.

API usage is a separate financial decision. If an application, automation pipeline or internal tool sends requests with an API key, those requests generate token-based charges independently of the ChatGPT Plus subscription. The published gpt-5.3-codex rates provide a starting point for estimates, but actual costs depend on input context, generated output, cached material, model selection and request volume.

What Plus users should remember

• Check the current usage dashboard because limits and allowances can change.
• Confirm Codex Cloud eligibility and workspace settings before planning around background execution.
• Treat the default cloud environment as a bounded development workspace, not production infrastructure.
• Review generated code, dependencies, tests and security-sensitive changes before release.