by OpenAI
OpenAI Codex review, pricing and verdict
Clones your repo, writes the code, opens the pull request. No keyboard required.
- Best for
- Backend engineers who want a multi-file refactor or a full pull request delegated rather than drafted line by line
- Not for
- Linux developers who want the Computer Use feature (macOS and Windows only)
Should you use OpenAI Codex?
Codex suits developers who want a multi-file change delegated end to end rather than typed line by line, replacing the job of reviewing AI-suggested diffs with reviewing finished pull requests. Treat its output as a draft from a fast junior engineer, not a merge-ready patch: every change still needs a human reviewer.
Because Codex ships as a plan feature inside ChatGPT rather than a bolt-on subscription, and the identical agent runs as a free, auditable CLI or a managed cloud sandbox, it removes both the extra bill and the vendor lock-in that rival coding agents ask a team to accept.
Bundling into the ChatGPT plan hundreds of millions of people already pay for, plus a foothold inside GitHub's own Agent HQ, gives Codex default distribution a standalone rival has to buy; but the clone-sandbox-PR workflow itself is not hard to replicate, and the open CLI carries essentially no lock-in on the client side.
Also for
- Engineering teams already paying for ChatGPT Plus or Pro who want a coding agent without a second subscription
- AppSec engineers screening pull requests for vulnerabilities through Codex Security
Show 3 more
- Knowledge workers outside engineering using Codex through Slack or the desktop app for lightweight automation
- Backend or platform engineers delegating multi-file refactors and PR generation
- AppSec engineers using Codex Security to triage vulnerabilities
Also not for
- Teams requiring a fully self-hosted or air-gapped coding agent
Main trade-off
Its July 2026 home-directory deletion bug is fixed, but it is a reminder that Codex's autonomy is only as safe as whichever sandbox and approval settings a given user actually has turned on.
OpenAI Codex pricing
- Top plan
- $500/mo
- Billing model
- Freemium
Codex has no subscription of its own: it is included in whichever ChatGPT plan a person already pays for, from Free ($0) and Go ($8) through Plus ($20), Pro at $100, $200 or $500 a month, Business ($20 per user a month billed annually, $25 billed monthly) and custom Enterprise and Edu pricing. Ultrafast mode needs the top Pro tier or a qualifying Enterprise or Edu agreement. Usage beyond a plan's included quota draws down credits billed per token: on OpenAI's published Codex rates GPT-6.1 Sol costs 50 credits per million input tokens, GPT-6.1 Terra 25 and GPT-6 Luna 2.5, so routing routine edits to Terra or Luna is the cheapest way to stretch an allowance. A separate OpenAI API key bills per token at standard API rates instead of through ChatGPT credits and skips the cloud features. The ChatGPT record carries the full plan table. Checked 2026-11-06 on OpenAI's Codex pricing page.
Before you commit
- Extra usage beyond a plan's included quota draws down per-token credits, and GPT-6.1 Sol credits cost far more than Luna's, so defaulting every task to the stronger model can burn a budget fast.
- Enterprise-grade controls (SCIM, EKM, dedicated audit logs) sit behind the custom-priced Enterprise/Education tier, not the per-seat Business plan.
What OpenAI Codex does on its own
- Scope of autonomy
- Semi-autonomous
- Human approval boundary
- Human in the loop
- Maturity
- Enterprise
Safety controls
- Runs sandboxedSupported
- Audit loggingSupported
- Human in the loopSupported
- Output filteringNot supported
What it can operate
- Web browsing
- Code execution
- File system
- Shell commands
- API calls
- Image generation
Where OpenAI Codex runs
- Deployment
- Cloud · Local · API
- Where it runs
- Web · Desktop · CLI · API · IDE plugin
- Memory
- Long term
- Context window
- 272K tokens
Models it runs on
- GPT-6 Astra
- GPT-6.1 Sol
- GPT-6.1 Terra
- GPT-6 Sol
- GPT-6 Luna
Systems it connects to
| System | Type |
|---|---|
| GitHub | Native |
| VS Code | Native |
| Slack | Native |
| GitHub Actions | Plugin |
| Linear | Native |
| GitLab | Native |
Languages it supports
- Python
- JavaScript
- TypeScript
- Go
- Rust
- Ruby
- Java
- C++
Benchmark scores
- SWE-bench Pro (GPT-6 Sol, Sept 2026)
- 64.6 / 100
- SWE-bench Pro (GPT-6 Luna, Sept 2026)
- 62.7 / 100
What OpenAI Codex does
OpenAI Codex is an autonomous coding agent that turns a plain-English task into a finished pull request: it works in its own sandbox, edits files, runs the tests and iterates on failures. It runs on OpenAI's GPT-6 models, including GPT-6.1 Sol and Terra, behind a sandbox-and-approval safety layer; its open-source CLI reached version 0.167.0 on November 6, 2026.
Autonomous PR delegation
Clones a GitHub repo into an isolated sandbox, writes the change across files, runs the project's own test suite, and opens a pull request without further input.
Cloud tasks with reusable environments
Tasks can start on a laptop, be steered from a phone, or run unattended in the cloud, with environments a team defines once and reuses.
OS-level sandbox with approval policies
A sandbox restricts what files and commands the agent can touch at the operating-system level, and a separate approval policy pauses it before anything outside that boundary.
Auto-review mode
A reviewer agent clears most out-of-sandbox approval pauses automatically; OpenAI says this now accounts for the majority of internal Codex Desktop token usage.
Show all 11 features
Code Review
Summaries and diff exploration in the ChatGPT desktop app, with an optional automatic first-pass review in the cloud for GitHub and GitLab changes.
Codex Security Cloud
Repository-wide vulnerability scans that run when asked or on a timer, investigate each finding, merge duplicates and draft the fix.
Ultrafast speed tier
Ultrafast streams up to 8x faster on the top two OpenAI models for Pro $500 and qualifying Enterprise accounts, and drains included usage 8x quicker (6x on purchased credits).
Voice-steered CLI with /agents view
Dictate a task into the terminal, then watch every delegated job in one /agents list; CLI 0.167.0 adds an agent command center that marks unread results.
Slack, Linear and GitHub cloud integrations
Teammates can tag @Codex in a Slack channel, hand it work from Linear, or delegate a GitHub issue or pull request to it with @codex.
Codex SDK
A TypeScript SDK lets a team embed the same agent in its own tools with structured output parsing and session resumption, with more languages planned.
Computer Use
In the ChatGPT desktop app on macOS and Windows, Codex can see and operate graphical apps for tasks that command-line tools cannot reach, in supported regions and subject to plan limits.
Strengths and limits
Strengths
- No separate subscription: Codex comes bundled with the ChatGPT plan a person may already pay for, instead of billing on top of an existing AI subscription (see the cost FAQ for exact tiers).
- The Codex CLI's source is public on github.com/openai/codex under Apache-2.0, letting a team audit or fork the client instead of trusting a closed binary.
- The cloud sandbox runs multiple tasks in parallel in the background, so a developer can queue several changes and review pull requests as they land instead of babysitting one session at a time.
- It reaches developers where they already work: the desktop app, the terminal, VS Code, an iPhone, Slack, Linear and GitHub.
Limits
- Every change still needs a human reviewer: Codex opens a pull request, it does not guarantee a correct one, so a developer has to review, not just merge, what it produces.
- Computer Use works only on macOS and Windows, only in supported regions, and is marked limited by plan, leaving Linux users without that capability.
- In July 2026, a bug let Codex delete real files in a user's home directory when running in full-access mode with the sandbox turned off; OpenAI has since added checks on delete targets.
- There is no self-hosting option: every mode, CLI included, depends on OpenAI's cloud models, so a team cannot run Codex fully air-gapped.
- Ultrafast, the speed tier worth paying for, is out of reach on cheaper plans and consumes included usage at 8x the Standard rate.
OpenAI Codex alternatives
| Agent | Where it runs | Price | Maturity | Checked |
|---|---|---|---|---|
| OpenAI Codex This page | Cloud · Your machine · API | Free plan | Enterprise | 6 Nov |
 ClinePick Codex for a managed cloud sandbox and a delegate-and-review workflow bundled into a ChatGPT plan; pick Cline for an open-source agent that runs inside your own editor with the model provider you choose. | Your machine · API | Free plan | Available | 27 Sep |
 JulesPick Codex if you're already inside the ChatGPT ecosystem and want Slack, Linear and GitHub access; pick Jules if your stack is Google's and you want a coding agent from that same vendor. | Cloud · API | Free plan | Available | 25 Oct |
 CosinePick Codex for breadth of surfaces (CLI, cloud, Slack, mobile, VS Code); pick Cosine for a smaller, coding-only vendor without a broader consumer chat platform attached. | Cloud · Your machine · API | $19/mo | Available | 18 Aug |
Also compared
Muse CodePick Codex if you want a coding agent bundled into a ChatGPT plan; pick Muse Code if your organization already standardizes on Meta's model family.
OpenAI Codex facts
- Works offline
- No
- Minimum skill level
- Intermediate
- Budget fit
- Mid
- Open source
- Yes
- Ships
- Weekly
- Last major update
- Maker founded
- 2025
- Headquarters
- San Francisco, CA
What changed
Questions about OpenAI Codex
What does OpenAI Codex cost in 2026?
Codex is included in a ChatGPT plan rather than billed separately: Free is $0 a month, Go is $8, Plus is $20, and Pro comes in $100, $200 and $500 tiers. Business is priced per user, cheaper on an annual contract, and Enterprise and Education pricing is negotiated directly. Usage past a plan's included quota is billed in credits.
Is there a free OpenAI Codex plan?
Yes: the Free and Go ChatGPT tiers include Codex with GPT-6 Luna at Standard speed in the desktop app, subject to rollout. Plus adds the web, CLI, IDE extension, iOS app, cloud integrations such as automatic code review and Slack, and GPT-6.1 Sol and Terra, so anyone doing more than small tasks will hit the wall and either wait or upgrade.
Which tools compete with OpenAI Codex?
The closest competitors are Claude Code from Anthropic, Jules from Google, Cline, Gemini CLI and Cosine, a smaller independent vendor. Jules and Gemini CLI fit teams standardized on Google's stack, Cline is an open-source agent that runs inside an editor, and Cosine suits a buyer who wants a focused coding vendor with no broader chat platform attached. Cursor is a different shape of product, an AI-native editor rather than a delegate-and-review agent, for teams that want to stay hands-on in the IDE.
Should I choose Codex or Claude Code?
Neither wins outright: Codex has the distribution edge, included in a ChatGPT plan that a very large user base already pays for, while Claude Code needs its own Anthropic subscription. Reviewers tend to find Codex quicker for greenfield work such as scaffolding a new API from nothing, while Claude Code earns praise for making fewer speculative edits on large, tangled refactors. Both ship as a terminal-first CLI with a cloud option, so the pick usually comes down to which model family a team trusts more.
How do I get started with OpenAI Codex?
Sign in with an existing ChatGPT account, since Codex uses that subscription rather than a separate signup. From there, install the Codex CLI from GitHub or add the Codex extension in VS Code, point it at a repository, and describe the task in plain English; a first run typically finishes inside the sandbox within a few minutes, after which you review the diff or pull request it opens.
About OpenAI Codex
OpenAI relaunched Codex in April 2025 as an agentic coding system, retiring the branding from its 2021 code-completion API. Built by OpenAI, it takes a plain-English task, clones the target GitHub repository into an isolated sandbox, and works through the change itself instead of only suggesting code for a human to accept line by line. The open-source CLI repository on GitHub shows 133.5k stars and an Apache-2.0 licence as of 6 November 2026.
Read the full background
Codex runs on the GPT-6 family OpenAI shipped on September 22, 2026. Its docs recommend GPT-6 Astra for the most demanding work, GPT-6.1 Sol for near-Astra performance at a lower cost, GPT-6.1 Terra as a faster, cheaper choice for writing, visual understanding and computer use, and GPT-6 Luna for focused, repeatable tasks; GPT-6 Sol stays available too. Which of them a person can pick depends on the plan. Anyone weighing GPT-6 against rival model families can compare the options in Which LLM Should You Use in 2026. A sandbox layer enforces what files the agent can touch and which commands it can run at the operating system level, while a separate approval policy decides when it must pause for a human; an Auto-review mode lets a second reviewer agent clear most of those pauses automatically. Planning a multi-file change, running the test suite, iterating on failures, and opening the pull request without a person retyping each step is what separates Codex from an autocomplete-style coding assistant.
It fits software teams that want a change delegated rather than drafted: point Codex at a task and it will scaffold a new API or landing page, refactor a legacy module, generate unit tests, or diagnose a performance regression while a developer works on something else. That is a different shape of help than Cursor, an AI-native editor built for staying hands-on in the IDE rather than handing a task off. Cloud integrations let a teammate tag @Codex in Slack or Linear, or hand it a GitHub issue or pull request, and ChatGPT Work, OpenAI's agent for non-coding knowledge work, shares Codex's usage allowance and credits. Hokai tracks it inside the coding-agent section of the directory alongside the picks in Best AI Coding Assistants in 2026.
There is no separate Codex subscription: it comes with the ChatGPT plan a person already has, from a $0 Free tier up through custom Enterprise and Edu pricing. Access comes through the ChatGPT desktop app and web, a Visual Studio Code extension, an iOS app, and the open-source Codex CLI. OpenAI's pricing page notes that Free and Go get only GPT-6 Luna, inside the desktop app, while the web, CLI, IDE extension, iOS and cloud integrations start at Plus.
OpenAI's DevDay on September 29, 2026 widened the product in four directions. Codex in the cloud runs on a computer, remotely from a phone, or from any device, and reusable development environments let a team start tasks quickly from a shared setup with approved settings and permissions. The refreshed CLI starts and steers tasks by voice and adds an /agents view for tracking several delegated tasks at once; the 0.167.0 release of November 6, 2026 adds an agent command center (codex agents --no-daemon) that marks unread results. A Code Review experience in the ChatGPT desktop app summarizes diffs and lets a person question Codex before commenting on GitHub pull requests or GitLab merge requests, and automatic reviews let Codex take a first pass in the cloud while the developer is away. Codex Security scans whole GitHub repositories in the cloud, checks new commits, removes duplicate findings and prepares fixes. Separately, the Agents API supports computer use and brings Codex's multi-agent capabilities into a developer's own application, covered in the Agents API pricing guide. OpenAI also sells an Ultrafast speed tier for its two strongest models.
Codex is the wrong pick in a few cases. Ultrafast is reserved for the top Pro tier and Enterprise, so most paying subscribers cannot buy it. Computer Use, which lets Codex operate graphical apps, now works on macOS and Windows but only in supported regions and with plan limits, there is no self-hosted option, and every change still needs a human reviewer. A team that wants to stay hands-on inside an editor will do better with Cursor, and one that prefers another model family can compare Claude Code first.
Codex ships changes to its CLI several times a week: version 0.167.0 was the latest stable release on November 6, 2026, and GPT-6.1 Terra needs CLI 0.159.0 or later. That pace has a downside: in July 2026, users running Codex in full-access mode with sandbox protections turned off hit a bug where the agent redirected its $HOME variable to build a temp directory and a faulty delete then removed real files in their home folder; OpenAI says it now verifies delete targets first and recommends staying in a sandbox mode (reported by The Decoder). OpenAI has also pushed Codex beyond pure coding, with a plugin system and Codex Security, a dedicated agent for finding and proposing fixes to vulnerabilities. Rivals move just as fast: in late July 2026, Windsurf was folded into Cognition and rebranded Devin Desktop, a reminder that the coding-agent field Codex competes in keeps consolidating.
Sources and guides
Official sources
- OpenAI Codex official site Vendor
- GitHub repository GitHub
Autonomy, guardrails, integrations and pricing on this page are read from OpenAI's own documentation. The verdict is Hokai's reading of them, not the vendor's claim.
Hokai guides covering OpenAI Codex
Also mentioned in
Lists that include OpenAI Codex
More AI Agents on Hokai
OpenHandsOpen-source AI agent that plans, codes, tests, and ships pull requestsPromptlessAI teammate that drafts doc updates as pull requests when you ship.
QualGentAutonomous QA agent that tests iOS and Android apps like a human.
OpenCodeOpen source coding agent that works with any model, in terminal, desktop or IDE.Codex SecurityOpenAI's security agent that scans repos and drafts fixes.
Antigravity CLIGoogle's terminal agent for multi-file coding, subagents and browser checks.