OpenAI Codex
OpenAI's coding agent across a CLI, an IDE extension, a cloud environment, ChatGPT and GitHub code review, powered by the GPT-5.6 family.
Updated 2026-08-02
Yes, within limits. OpenAI Codex runs a free tier with paid plans above it.
Listed pricing: Included with ChatGPT Free, Go, Plus, Pro, Business, Edu and Enterprise plans.
Overview
Codex is OpenAI's coding agent, and the framing in its own documentation is deliberately loose: "Start with a goal, idea, or task. ChatGPT can gather context, take action, and produce something useful." What makes it concrete is the surface list. There is a Codex CLI for the terminal, a Codex IDE extension, Codex cloud for hosted runs, the ChatGPT desktop and web apps, and a GitHub code review integration that comments on pull requests. AGENTS.md is the configuration file a repo checks in to tell the agent how to behave, and sandboxing plus MCP support are documented parts of the product.
The models underneath are the GPT-5.6 family, published as three variants with different jobs. Sol is the capability tier, Terra is the balanced one, Luna is for speed and lower cost. That split matters more than it looks, because Codex's usage accounting is model-dependent: the same plan buys wildly different amounts of work depending on which variant a task routes to.
My read: the bundling is the strategy. OpenAI's own documentation says Codex "is included in your ChatGPT Free, Go, Plus, Pro, Business, Edu, or Enterprise plan", which means an organisation that already bought ChatGPT seats has already bought a coding agent, and the procurement conversation for a competitor starts from behind. The catch is that the published limits are ranges rather than numbers. OpenAI expresses them as messages per rolling five-hour window, and the range spans two orders of magnitude depending on model choice, so the honest answer to "how much do I get" is that nobody can tell you before you run it. Overflow is sold as credits priced per million tokens, which at least makes the marginal cost legible even when the included allowance is not.
Is OpenAI Codex free?
Yes, within limits. OpenAI Codex runs a free tier with paid plans above it.
What the free tier covers: Yes. OpenAI documents Codex as included on the ChatGPT Free plan, with access described as restricted to basic tasks.
Listed pricing: Included with ChatGPT Free, Go, Plus, Pro, Business, Edu and Enterprise plans.
Pricing verified . That is the date the plans were last re-checked against the vendor's own pages, not today's date. Confirm on the official site before you pay.
What the free tier leaves out
Read straight off the plan list below. Vendors move features between tiers, so check the current split before you pay.
- ChatGPT Plus $20/mo Documented as roughly 10 to 2,000 messages per rolling five-hour window depending on which GPT-5.6 variant the work routes to.
- ChatGPT Pro Higher-priced individual tier with 5x and 20x usage levels Larger message allowances per five-hour window and higher rate limits. Confirm the current price on OpenAI's pricing page before buying.
- ChatGPT Business $20/user/mo Documented as matching the Plus allowance per five-hour window, with workspace administration on top.
- Edu and Enterprise Custom Codex is listed as included. Enterprise plans on flexible pricing can purchase additional workspace credits.
- Additional credits 125 credits per million input tokens on GPT-5.6 Sol, 5 credits per million on Luna Bought on top of any plan so work continues past the included allowance without a tier upgrade.
OpenAI Codex pricing
| Plan | Price | What's included |
|---|---|---|
| ChatGPT Free and Go | $0 and the Go tier price | OpenAI lists Codex as included on both, with more restricted access than the paid tiers. |
| ChatGPT Plus | $20/mo | Documented as roughly 10 to 2,000 messages per rolling five-hour window depending on which GPT-5.6 variant the work routes to. |
| ChatGPT Pro | Higher-priced individual tier with 5x and 20x usage levels | Larger message allowances per five-hour window and higher rate limits. Confirm the current price on OpenAI's pricing page before buying. |
| ChatGPT Business | $20/user/mo | Documented as matching the Plus allowance per five-hour window, with workspace administration on top. |
| Edu and Enterprise | Custom | Codex is listed as included. Enterprise plans on flexible pricing can purchase additional workspace credits. |
| Additional credits | 125 credits per million input tokens on GPT-5.6 Sol, 5 credits per million on Luna | Bought on top of any plan so work continues past the included allowance without a tier upgrade. |
OpenAI lists Codex as included on both, with more restricted access than the paid tiers.
Documented as roughly 10 to 2,000 messages per rolling five-hour window depending on which GPT-5.6 variant the work routes to.
Larger message allowances per five-hour window and higher rate limits. Confirm the current price on OpenAI's pricing page before buying.
Documented as matching the Plus allowance per five-hour window, with workspace administration on top.
Codex is listed as included. Enterprise plans on flexible pricing can purchase additional workspace credits.
Bought on top of any plan so work continues past the included allowance without a tier upgrade.
One plan carries no published number: Edu and Enterprise, recorded as “Custom”. Nothing is estimated in its place. The vendor's own pricing page is the only source for those figures.
Pricing verified . That is the date the plans were last re-checked against the vendor's own pages, not today's date. Confirm on the official site before you pay.
Is OpenAI Codex worth it?
Worth it for Teams already standardised on ChatGPT who want the coding agent bundled with the seats they pay for, and who want the same agent in a terminal, an editor and a pull request.
You can test that on the free tier before paying anything. The recorded trade-offs are listed below, and any one of them can settle the question on its own.
The 8.8/10 AI Score is an editorial read of published capability, price and shipping pace. Nobody here has hands-on hours with OpenAI Codex. How we verify.
Worth it if
The strengths recorded against this entry.
- Included in ChatGPT plans from the free tier upward, so most organisations already own it
- Covers the terminal, the editor, a hosted cloud runner, the ChatGPT apps and GitHub pull request review from one entitlement
- AGENTS.md keeps agent configuration in the repo and under version control rather than in someone's local settings
- Three GPT-5.6 variants let a team route cheap work to Luna and hard work to Sol instead of paying one rate for everything
- Overflow credits are priced per million tokens, which makes the marginal cost of extra work legible
Not worth it if
Any one of these blocks your use case.
- Published limits are ranges per rolling five-hour window that span two orders of magnitude, so the included allowance is not knowable in advance
- Usage depends on which model variant a task routes to, and routing is not fully under the user's control
- Pro pricing is presented as tiers rather than a single figure, so the real cost of the heavy plan needs checking at the point of sale
- The agent is tied to a ChatGPT account, which is awkward for teams that standardised on a different assistant vendor
What sets OpenAI Codex apart
- One agent spanning a CLI, an IDE extension, a hosted cloud environment, the ChatGPT apps and GitHub pull request review
- Runs on the GPT-5.6 family, split into Sol for capability, Terra for balanced work and Luna for speed and lower cost
- AGENTS.md gives a repo a checked-in configuration file the agent reads
- Sandboxing and MCP support are documented parts of the product rather than add-ons
- Bundled into ChatGPT plans down to the free tier instead of sold as a separate seat
Key features
Codex CLI
The terminal surface, for driving the agent against a local repo without leaving the shell. It shares the same account and entitlement as the other surfaces.
Codex IDE extension
Puts the agent inside the editor rather than beside it, for developers who want review and acceptance of changes in the same window where they read the code.
Codex cloud
A hosted environment for runs that should not tie up a laptop. Combined with the CLI and IDE surfaces, the same task can be started in one place and picked up in another.
GitHub code review
Reviews pull requests directly, which is the surface that reaches reviewers who never install anything. It is also the one that puts agent output in front of the whole team rather than one developer.
AGENTS.md, sandboxing and MCP
AGENTS.md is a checked-in file that configures how the agent works in a given repo, so behaviour is versioned alongside the code. Sandboxing bounds what a run can touch, and MCP support connects it to outside tools.
GPT-5.6 model split
Sol for capability, Terra for balanced work, Luna for speed and lower cost. Usage limits and credit prices both vary by variant, so model selection is a budgeting decision as much as a quality one.
How it compares
| Tool | Best for | Pricing | Score |
|---|---|---|---|
| OpenAI Codex | Teams already standardised on ChatGPT who want the coding agent bundled with the seats they pay for, and who want the same agent in a terminal, an editor and a pull request. | Included with ChatGPT Free, Go, Plus, Pro, Business, Edu and Enterprise plans | 8.8/10 |
| Claude Code vs Claude Code → | Developers who want an agent that works inside an existing repo and toolchain rather than in a hosted editor, and who already pay for a Claude plan. | Included with Claude Free, Pro, Max, Team and Enterprise plans | 9.3/10 |
| Cursor vs Cursor → | Professional developers handling complex, multi-file refactors who want AI built into a familiar VS Code-based editor. | Freemium | 9.5/10 |
| GitHub Copilot vs GitHub Copilot → | Developers who want an AI coding assistant built into their existing editor, from free-tier autocomplete to autonomous PR workflows. | Free tier + Pro $10/mo | 8.8/10 |
| Devin | Engineering teams that want to hand off well-scoped coding tasks, like migrations, bug fixes, and boilerplate, to an autonomous agent working independently. | Teams ~$500/user/mo, Enterprise custom | 8.5/10 |
Compare head-to-head
Related reading
OpenAI Codex Thursday: Goal Mode & Appshots
OpenAI's May 22 Codex update adds always-on Goal mode, Appshots for instant app context, and remote Mac control while locked.
OpenAI Codex Goes Mobile: ChatGPT App Preview
OpenAI added Codex to the ChatGPT mobile app on May 14, letting developers start and steer coding agents from iOS and Android across all plans.
Claude Eval Escapes: An 8-Day Disclosure Clock
Anthropic suspended the cyber evals July 23, notified affected organizations July 27, published July 31. What those dates settle and what they don't.
Ready to try OpenAI Codex?
Head to the official site to start with OpenAI Codex — pricing and plans are listed above.
Visit OpenAI Codex