OpenAI is renewing its Codex for OSS program, and the headline number is 10,000 grants, double the original 5,000. What is not settled is which ChatGPT Pro tier those grants carry. The person who announced the renewal wrote "$100 Pro plans." An aggregator summary of the same news says the $200 tier. OpenAI's own program page, as we fetched it, gives neither figure.
This post sorts out what is claimed, what is confirmed, and how a maintainer can check for themselves. If you maintain an open-source project, it also covers how to apply. For wider context on ChatGPT's tiers, see our coverage of OpenAI's paid usage resets on the $100 and $200 plans and the true cost of an AI subscription.
TL;DR: claimed vs verified
| Claim | Who said it | Status |
|---|---|---|
| Program renewed six months after launch | Charlie Marsh on X | Stated by the announcer; consistent with the program's March 2026 launch |
| Grants doubled from 5,000 to 10,000 | Same post | Stated by the announcer; not on the OpenAI program page we fetched |
| Recipients get $100 Pro plans | Same post | Stated by the announcer |
| Recipients get the $200 Pro plan | Aggregator summary, citing a later post | Conflicts with the original wording |
| Past recipients upgraded automatically | Aggregator summary | Conflicts with "we encourage you to re-apply" |
| Six months of ChatGPT Pro with Codex, Codex Security access, API credits | OpenAI Codex for Open Source page | Confirmed by OpenAI docs |
What OpenAI's own page confirms
The official Codex for Open Source page says maintainers can apply for API credits, six months of ChatGPT Pro with Codex, and Codex Security. It explains that over the past year the Codex Open Source Fund, described as a $1 million fund, has supported projects that needed API credits, including teams using Codex to power GitHub pull request workflows. The fund now supports eligible maintainers by offering six months of ChatGPT Pro with Codex and conditional access to Codex Security for core maintainers with write access.
The page lists what the program includes:
- Six months of ChatGPT Pro with Codex for day-to-day coding, triage, review and maintainer workflows.
- Conditional access to Codex Security for repositories that need deeper security coverage, reviewed case by case.
- API credits through the Codex Open Source Fund for projects that use Codex in pull request review, maintainer automation or release workflows.
It also says developers should code in the tools they prefer, naming Codex, OpenCode, Cline, pi and OpenClaw, and that the program supports that work. Who should apply: core maintainers and people who run widely used public projects, plus projects that do not fit the criteria but play an important role in the ecosystem, who should apply anyway and explain why.
What the page does not say, in the version we fetched, is a grant count or a plan price. That is why the 10,000 figure and the $100 versus $200 question must be read as announcements on social media, not documentation.
Where the $100 versus $200 confusion comes from
OpenAI sells several ChatGPT tiers. As we documented in our usage-reset post, a $100 Pro Lite tier sits beside the $200 Pro plan. A grant "at the $100 level" and one "at the $200 level" are therefore different products, and the difference matters: the higher tier carries more usage.
Charlie Marsh, who announced the renewal, wrote: "We started Codex for OSS six months ago. Today, we're renewing the program with $100 Pro plans and doubling the number of grants (from 5,000 to 10,000) to reach more maintainers. (If you received a grant previously, we encourage you to re-apply!)"
An aggregator summary published a day later reads differently. It describes 10,000 subscriptions to the $200 Pro tier, says the plan value rises from "the previous $100 level," and says prior holders are upgraded automatically. It quotes a follow-up post saying there are "still 10,000 grants in this round, now at the higher tier."
Possible explanations, none confirmed:
- The program first launched on the $100 tier and the renewal moved to $200 (consistent with the aggregator, inconsistent with Marsh's wording).
- The renewal is at $100 and the aggregator mixed up tiers.
- OpenAI changed the tier between the first announcement and the follow-up.
We could not resolve this from public pages. The practical answer is to read your approval email, which should name the plan, and to treat any tier as subject to OpenAI's discretion: the program terms state that benefits are determined by OpenAI in its sole discretion.
Why this matters for maintainers
Open-source release shipping out, representing Codex for OSS grants for maintainers
Open-source maintainers are the group hit hardest by the AI era's two-sided pressure: they are asked to review a rising volume of AI-generated contributions, and they rarely have budget for premium tools. We have tracked the volume side in Google pausing its open-source bug bounty over AI slop reports and the policy side in Linus Torvalds on AI and kernel review. A grant of a paid plan gives maintainers the same class of tools the contributors use.
Competition also matters. Anthropic runs a comparable program; see Claude for Open Source, expanded to Max 20x. The two labs are, in effect, competing for the habits of the people who maintain critical infrastructure. For a maintainer, that is leverage: you can apply to both and pick the tool that fits your workflow.
What the free plan will and will not fix
- It helps with: triage, review, release notes, test generation and refactors, using higher usage limits than a free account.
- It does not fix: the review burden from low-quality AI-generated pull requests, which no plan upgrade removes.
- Time limit: six months, per the program page. Plan for what happens after, including whether your project can fund a seat.
- Data and terms: applying means agreeing to the program terms. Read them for how your usage data and repositories are handled before enabling Codex Security on private or sensitive code.
A short timeline of the program
- Early March 2026: OpenAI launches Codex for Open Source with API credits, free ChatGPT Pro and conditional Codex Security access for maintainers, as covered by outlets such as Winbuzzer.
- Spring and summer 2026: the program is described by third-party guides as covering API credit grants for projects using Codex in pull request workflows. Those guides cite figures such as a per-project credit cap that we could not find on OpenAI's own page, so we do not repeat it as fact.
- October 2026: Marsh announces the renewal, saying the program has run for six months and that the grant count is doubling.
The pattern is typical of lab programs: a launch with a fixed cohort, a half-year term, then a renewal with a larger cohort. The unusual part this time is the ambiguity about the tier, which is why we recommend relying on your own approval email.
Decision guide for maintainers
| Your situation | What to do |
|---|---|
| Never applied, maintain a widely used project | Apply now; rolling review means earlier is better |
| Held a grant in the first round | Re-apply as Marsh advises, and also check your account in case an upgrade is applied |
| Small project, not obviously critical | Apply anyway and explain the ecosystem role, as the page invites |
| Need security scanning on private code | Request Codex Security explicitly and expect case-by-case review |
| Already using a competing lab's program | You can hold both; compare limits before committing workflows to one |
How to apply or re-apply
- Open the Codex for Open Source page and use the application link.
- Describe the project, its usage and why it is important to the ecosystem. The criteria named by OpenAI include repository usage, ecosystem importance, evidence of active maintenance and your role or permissions.
- State which benefits you need (Pro seat, API credits, Codex Security).
- If you had a grant before, follow Marsh's guidance and re-apply rather than assuming an automatic upgrade.
- Watch your email: applications are reviewed on a rolling basis and selected applicants are notified by email.
How to verify the plan you received
- Check the plan name in ChatGPT billing settings once the grant is applied.
- Compare usage limits against OpenAI's published limits for the $100 and $200 tiers.
- Keep the approval email: it is the only document that states your tier and end date.
If you find the program page or your email contradicts what we wrote, treat the primary source as correct and let us know through @explainx_ai.
What to watch next
- Whether OpenAI updates the program page with the grant count and tier.
- Whether Codex Security access widens beyond core maintainers with write access.
- Whether competing programs match the 10,000 number.
- Whether the six-month term is extended for projects that renew.
Related reading
- OpenAI paid usage resets on the $200 Pro plan
- What your $20 AI subscription actually costs
- Claude for Open Source, expanded
- Google pauses open-source bug bounty over AI slop
- Linus Torvalds on AI and kernel review
- Codex Auto-Review is free
- Official: Codex for Open Source
Details are accurate as of October 22, 2026. Program terms, grant counts and plan tiers can change; confirm with OpenAI.