OpenComputer vs Codex cloud

Codex cloud is OpenAI's own way to run Codex in the cloud: each task runs in its own VM inside ChatGPT, started from an environment Codex prepares from your GitHub repositories, and it is included in paid ChatGPT plans. OpenComputer runs the same Codex agent (and Claude Code) on a cloud Linux machine per Slack workspace, driven from a shared Slack thread or an API, and billed per Slack workspace for work minutes.

OpenComputer is not a replacement for Codex. It runs Codex, signed in with your own ChatGPT subscription if you like. The difference is everything around the agent: who can hand it work, where that happens, what stays warm between tasks, and how it is billed.

If you are one developer already paying for ChatGPT Plus or Pro, Codex cloud is the simpler pick: there is no separate charge for the VM. If a team wants to hand work to an agent in Slack, steer it together and pay per workspace, OpenComputer fits.

At a glance

What is Codex cloud?

Codex cloud is the cloud side of OpenAI's coding agent, found under Work in > Cloud in ChatGPT. You choose GitHub repositories and Codex inspects them, installs dependencies and tools, and tests the workflow, then saves that as an environment you review and publish. Each new task gets its own VM from the published environment. When the task is done you review the changed files and test results, ask for follow-ups, and commit or open a pull request when you are ready.

You start tasks from ChatGPT on the web or desktop, Codex in the ChatGPT mobile app, or the terminal with codex cloud exec, and codex apply brings a cloud diff to your local checkout. Environments can reach private networks over Tailscale and use short-lived cloud credentials through OIDC. Codex cloud uses OpenAI models only and requires a ChatGPT sign-in: it is unavailable with an API key.

What is OpenComputer?

OpenComputer runs Claude Code and Codex in its own cloud, and you hand them work by mentioning @Computer in a Slack thread or calling the API from your code. Each Slack workspace gets its own cloud Linux machine, sized to its plan. Every thread works in its own copy of the repos on its own branch, so parallel requests don't collide. Workspaces resume warm, with the repository, dependencies and build caches already in place.

@Computer reads the code, runs the commands and tests, then opens a pull request that says what changed, why, and how it was checked, with a link back to the thread. UI changes come with a screenshot and a preview link. OpenComputer has no agent harness of its own: the agent doing the work is Codex or Claude Code.

Key differences

Same agent, different cloud

Both products run Codex on a cloud machine. With Codex cloud, OpenAI runs a VM per task, the environment and the agent, and you get the closest fit with the rest of Codex: the same CLI, codex apply, repository skills and OpenAI's latest models. With OpenComputer, each Slack workspace gets one cloud machine where every thread works in its own copy of the repos on its own branch, and the agent can be Codex or Claude Code, per request. If you want every task isolated in its own VM, Codex cloud gives you that.

Who hands it work, and where

Codex cloud is centered on the ChatGPT apps and CLI, organized around your own tasks. Delegating from Slack or Microsoft Teams with @ChatGPT is for Enterprise workspaces and needs a workspace-shared published environment. On OpenComputer, Slack is the main interface on every plan, and the thread is the unit of work: everyone in it sees every file @Computer opens and every command it runs, anyone can steer it, and conventions one person explains apply to later requests. The thread records who asked for each change and why.

Environments and what persists

Codex cloud's environment setup is agent-driven: Codex prepares the environment, tests it and asks you to publish it. Each task then keeps its own files, including uncommitted changes, and its saved VM state is recoverable for up to seven days after you last start a turn or resume it. VM size depends on plan: 2 vCPU and 8 GiB on Plus, 4 vCPU and 16 GiB on Pro, Business and Enterprise. OpenComputer keeps each workspace's repository, installed dependencies and build caches on its machine, so workspaces resume warm without a setup step. Codex cloud documents private network access over Tailscale, which OpenComputer does not list.

Models and billing

Codex cloud tasks share your ChatGPT plan's usage with local Codex messages and may use more of it. Beyond the limits you buy credits. OpenComputer can use the same ChatGPT subscription, a Claude subscription for Claude Code, your own API key, or its own tokens at API prices with no markup. When a subscription hits its limit you choose whether it switches to OpenComputer tokens until the reset or stops and waits. You pay OpenComputer separately for work minutes.

Pricing compared

As of October 2026, OpenAI lists ChatGPT Plus at $20 a month with Codex cloud, Pro from $100 a month (with $100, $200 and $500 options), and Business at $25 per user per month billed monthly or $20 billed annually. Cloud is not listed on Free or Go. OpenComputer lists Pro at $20 a month with 2,000 work minutes and Max at $200 with 25,000, for the whole Slack workspace, with $0.01 per extra minute.

For one developer on ChatGPT Plus, Codex cloud adds $0. Using OpenComputer with that same subscription adds $20 a month for OpenComputer Pro, so OpenAI's own cloud is cheaper. For a team of five, ChatGPT Business costs $125 a month billed monthly and gives each person their own Codex. OpenComputer Pro is $20 a month for the whole workspace, covering about 33 hours of active agent work, plus the model: a ChatGPT or Claude subscription you already have, an API key, or OpenComputer tokens at API prices. The units differ: ChatGPT plans meter usage per seat, while OpenComputer meters the minutes the agent spends working.

When to choose Codex cloud

• You are one developer, already on ChatGPT Plus or Pro, and want cloud tasks at no extra cost.
• You want Codex to prepare and test the environment for you, each task in its own VM, and task state kept for up to seven days.
• You need private network access over Tailscale or short-lived cloud credentials through OIDC.
• You use the ChatGPT apps and Codex CLI and want cloud tasks next to your local ones.

When to choose OpenComputer

• Your team works in Slack and wants one shared thread where anyone can see and steer the agent, without an Enterprise plan.
• You want both Codex and Claude Code available from the same place.
• You want pull requests with the evidence attached, including a screenshot and preview link for UI work.
• You'd rather pay per workspace for active agent time than add seats, or you want to build your own serverless agents on the same cloud.

FAQ

Is OpenComputer a Codex cloud alternative?

Yes, in the sense that both run Codex on a cloud machine so it keeps working without your laptop. OpenComputer runs the same Codex agent, plus Claude Code, on its own cloud, driven from a shared Slack thread or an API. It is not a different coding agent.

Can OpenComputer use my ChatGPT subscription?

Yes. You can sign in with your own ChatGPT or Claude subscription and pay $0 to OpenComputer for the model, use your own API key, or use OpenComputer's tokens at API prices with no markup. Work minutes are billed separately per Slack workspace.

How much does Codex cloud cost?

There is no separate VM charge. Codex cloud is included in ChatGPT Plus at $20 a month, Pro from $100 a month, and Business at $25 per user per month billed monthly, and it shares the plan's usage limits. You can buy credits for usage beyond the limits.

Can I use Codex cloud with an OpenAI API key?

No. An API key works for the local Codex CLI, SDK and IDE use, but Codex cloud requires a ChatGPT sign-in. OpenComputer accepts either a ChatGPT subscription or your own API key for Codex.

Does Codex cloud work in Slack?

Yes, on Enterprise workspaces: you mention @ChatGPT in Slack or Microsoft Teams with a workspace-shared environment, and the result comes back to the thread. OpenComputer works in Slack on every plan, with the whole thread able to see and steer the work.

Sources

• Codex cloud
• Cloud environments
• Codex pricing
• Codex in Slack
• Developer commands