OpenAI has made Codex availability easier to evaluate for engineering teams. In an official Help Center article visible on June 1, 2026, OpenAI says Codex is included across Free, Go, Plus, Pro, Business, Edu, and Enterprise plans. The same article also says Codex usage limits depend on plan and count toward a user’s broader agentic usage allowance. Alongside OpenAI’s recent Codex updates on May 14, 2026 for mobile access and May 29, 2026 for Windows computer use, this gives QA leaders a clearer picture of where Codex now fits inside real team workflows.
This is a documentation-led update rather than a brand-new model launch, but it still matters. Many QA teams have been testing AI coding assistants informally. Clearer official guidance on plan coverage, usage limits, and admin controls makes it easier to decide whether Codex can move from individual experimentation into a governed testing workflow.
What OpenAI now says about Codex access
Across OpenAI’s current Codex documentation, three points stand out for QA and SDET teams:
- Codex is no longer positioned as a narrow premium-only tool: OpenAI says it is included across ChatGPT Free, Go, Plus, Pro, Business, Edu, and Enterprise plans, although access levels vary.
- Usage is tiered: OpenAI’s pricing documentation shows that Free and Go have limited trial-style access, while Plus and Pro offer larger usage windows and more complex coding sessions.
- Governance features differ by workspace type: the Help Center article points Business and Enterprise or Edu users to workspace app controls, role-based access control, and Compliance API coverage for supported Codex clients.
That combination is useful because rollout questions are no longer just about model quality. They are also about who can access Codex, how much usage a plan supports, and what oversight a QA organization can keep when the tool is used in testing, debugging, and code review.
Why this matters for QA engineers
For QA engineers, the practical signal is simple: Codex is becoming easier to pilot across mixed team sizes without starting from an API-only or high-cost enterprise assumption. That lowers the barrier for test engineers who want to try AI-assisted defect triage, automation review, flaky-test investigation, and small framework changes inside the same family of ChatGPT plans their organizations may already use.
- Smaller QA teams can pilot earlier: limited Free or Go access may be enough to evaluate prompt quality, code review usefulness, or log analysis workflows before committing to a wider rollout.
- Larger automation teams can plan governance sooner: Business, Edu, and Enterprise guidance around app controls, RBAC, and compliance logging helps teams map Codex to existing security and audit requirements.
- Usage planning becomes more concrete: because Codex counts toward agentic usage limits, teams need to separate lightweight prompt checks from heavier codebase sessions and long-running tasks.
Three rollout checks QA teams should make this week
- Confirm plan coverage: check whether your current ChatGPT plan already includes enough Codex access for the kind of testing work you want to pilot.
- Define approved use cases: start with bounded tasks such as summarizing failed runs, reviewing generated tests, or proposing locator fixes before allowing broader repository changes.
- Track cost and auditability: if engineers start using Codex heavily for large repos or long sessions, monitor how quickly usage windows are consumed and whether compliance logging meets your review needs.
Recent context from OpenAI’s May updates
This clarification lands after two other recent Codex updates from OpenAI. On May 14, 2026, OpenAI said Codex entered preview in the ChatGPT mobile app across all plans, including Free and Go. On May 29, 2026, OpenAI added Windows computer use and Codex profiles in ChatGPT release notes. Read together, those updates suggest OpenAI is not only improving Codex features but also making its access model clearer for broader day-to-day engineering use.