The open-source maintainer sits at the center of the modern software ecosystem, with the libraries they build often deep within production systems. That influence has made them attractive to AI companies trying to win developers over to their coding tools.OpenAI announced on Friday that it was accepting applications to join its own maintainer program, following Anthropic’s launch last week. (And there’s nothing in the fine print for either company that says you can’t be part of both programs.)OpenAI states that members of its Codex for Open Source program receive six months of ChatGPT Pro with Codex, conditional access to Codex Security for repositories that need deeper security coverage, API credits through the Codex Open Source Fund for projects that use Codex in pull request review, maintainer automation, release workflows, or other core OSS work.OpenAI’s announcement builds on the Codex Open Source Fund, which has already provided $1 million in API credits to open-source projects over the past year, OpenAI tells The New Stack. With this expansion, OpenAI is extending support directly to maintainers working on widely used infrastructure projects.Anthropic beat OpenAI by a week with the launch of its own Claude for the Open Source Program, which offers six months of free access to its premium Claude Max subscription for up to 10,000 open source maintainers.For Anthropic and OpenAI, the programs are far from entirely altruistic — they’re looking to get their coding assistants and models directly into the hands of developers who shape the tools many other engineers rely on.That is why neither program is open to just anyone: For the Anthropic program, applicants need either at least 5K GitHub stars or one million monthly npm downloads, along with evidence of active development in the previous three months.Meanwhile, the OpenAI program is a little more open-ended in its asks: “Maintainers of active open-source projects can apply,” and “we look for projects with meaningful usage, broad adoption, or clear importance to the software ecosystem” are among the more open-ended requirements.These filters — though OpenAI’s language is less obvious about what it’s looking for — will likely help the company select maintainers whose libraries are already widely used.Anthropic describes the program as its “way of saying thank you” to the developers who maintain popular software libraries. The sentiment offered by OpenAI is nearly identical: “OpenAI is grateful to the maintainers who keep that work moving.”While open source maintainers often work without direct funding, their projects underpin critical infrastructure across the software industry. Access to Claude Max — which normally allows hundreds of messages over a few hours and would typically cost about $1,200 over six months — could therefore prove invaluable for maintainers reviewing pull requests, refactoring large repositories, or navigating complex codebases. ChatGPT Pro costs about the same — $200 per month.“A credibility signal”Several maintainers have already confirmed that they’re participating in Anthropic’s program. Rostislav Dugin, one of the core contributors to Databasus, said the project had been accepted into Claude for Open Source earlier this week.Databasus is a free, open source backup management tool for databases, with a primary focus on PostgreSQL. The project has around 5.8K GitHub stars, more than 250,000 Docker pulls, and tens of thousands of daily active users.While Claude Max access is obviously beneficial for cash-strapped maintainers, Dugin said the recognition from Anthropic is just as important.“For me, this was a meaningful moment — not because of the subscription itself, but because of what it represents,” Dugin writes in a blog post. “A company like Anthropic looked at the project, evaluated it against their criteria, and decided it’s worth supporting. That’s a credibility signal that matters.”Elsewhere, Dax Raad, one of the developers behind the super-popular open source AI coding agent OpenCode, also appears to have been accepted into the program. That may raise a few eyebrows, given that Anthropic recently moved to restrict the use of Claude subscriptions through third-party coding tools such as OpenCode, part of a broader effort to steer developers toward its own tooling, such as Claude Code.Still, developers like Raad are exactly the type of maintainers that Anthropic is trying to reach. OpenCode has emerged as one of the fastest-growing commercial open-source projects, with 117K stars to date.OpenAI has similar goals for its program, and said as much in its announcement:“If you’re a core maintainer and are building a tool that our ecosystem depends on, we would love to support your work,” states OpenAI in a video released on Friday.“Oh, so that’s your little plan.”The announcements, barely a week apart, by OpenAI and Anthropic, arrive amid a broader effort by AI companies to win over developers whose work underpins much of today’s software stack. OpenAI runs the Codex Open Source Fund, which provides API credits to open-source projects. The fund has a total allocation of about $1 million, with grants of up to $25,000 per project.Google is coming at things from a slightly different angle. Gemini CLI — an open-source AI agent that runs in the terminal — ships with a fairly generous free tier that allows developers to run up to 1,000 model requests per day and as many as 60 requests per minute at no cost. That level of access means many developers can experiment quite substantively with Gemini gratis, which might go some way toward seeding its tools inside developer environments. (Google has not announced a similar targeted program.)Of course, Anthropic and OpenAI’s plans are more focused. Instead of offering broad free access, Claude for Open Source targets a narrower group of maintainers whose projects already see significant adoption.If the Anthropic program reaches its full allocation of 10,000 developers, that will be the equivalent of around $12 million in retail value. For a company that just raised $30 billion, the cost is modest. What it buys is something harder to price: access to the maintainers whose libraries power large parts of the software industry. (OpenAI hasn’t capped the number of people for its program and says it will review applicants on a rolling basis.)Whether Anthropic and OpenAI’s dueling programs ultimately translate into long-term users remains to be seen.Whether Anthropic and OpenAI’s dueling programs ultimately translate into long-term users remains to be seen. For now, it signals how intensely AI companies are competing for the developers who help shape the software ecosystem — even as some parts of the open-source community remain uneasy about how their work has been used to train AI models. That tension has resurfaced more recently, as maintainers report dealing with large volumes of “AI slop” infesting public repositories.However you slice and dice it, there’s no escaping the simple fact that what OpenAI and Anthropic are offering here is temporary. The Claude Max subscriptions last for six months. The same goes for ChatGPT Pro.In the words of Homer J. Simpson, who, after wolfing down the free samples, was invited to buy some cookies: “Oh, so that’s your little plan. Get us addicted, then jack up the price. Well, you win.”The post Anthropic and OpenAI are battling for the best open-source maintainers appeared first on The New Stack.