GitHub: "Generative AI is now the standard"

GitHub publishes current figures on its developer platform in the Octoverse report. It is growing – also thanks to AI support in programming.

listen Print view
Promotional image GitHub

(Image: Github)

4 min. read

GitHub is growing. According to its figures, the developer platform now has more than 180 million users and hosts 630 million software projects. Over the past twelve months, more than 36 million users have been added – also thanks to AI, GitHub believes.

The figures come from GitHub's annual Octoverse report. The Microsoft subsidiary presented it on October 28, 2025, at its developer conference GitHub Universe in San Francisco. Every minute, 230 new repositories are created, and 43 million change proposals are accepted monthly. In total, there were already around one billion commits in 2025.

According to the report, GitHub also owes its growth to its Copilot. Almost 80 percent of new users use GitHub's AI coding assistant in the first week. “Generative AI is now the standard in development,” the company writes. It also attracts developers: Six of the ten most active open-source projects are AI tools such as Codex or Ollama.

Videos by heise

Users also use Copilot for code review. Almost 73 percent of developers surveyed by GitHub stated that they rate their effectiveness higher with it. However, the figures should be treated with caution; a study by METR from the summer suggests that such anecdotal self-assessments are too high.

After Python took the top spot for the first time in 2024, TypeScript is now moving into first place. In August 2025, the scripting language was used more than Python and JavaScript for the first time. GitHub attributes this shift to the increasing use of AI-assisted programming. “The rise illustrates how developers are increasingly switching to typed languages that make agent-assisted coding more reliable in production,” writes GitHub.

According to GitHub, there are connections between the spread of AI tools and changing language preferences: “These and other changes indicate that AI is not only influencing how quickly code is written, but also which languages and tools developers use.” GitHub's AI coding app Spark, announced at the last Universe conference, also relies on TypeScript. Spark generates web applications based on text-based prompts in natural language.

The focus on AI is also reflected in GitHub's own product development. The company announced “Agent HQ,” the ability to use external AI agents from Anthropic, OpenAI, Google, xAI, and others in GitHub projects. The feature is available to paying Copilot subscribers starting this week.

Whether AI-assisted or not, what is on GitHub is only partially public. Private repositories account for the majority of source code hosted on GitHub, at around 82 percent. In total, there are almost 5 billion projects used by paid enterprise and team accounts.

GitHub nonetheless emphasizes the importance of open-source projects. Although they accounted for a smaller share, they “provide the libraries, models, and workflows that drive the entire ecosystem.” Germany, incidentally, ranks 5th among countries of origin with the most contributors to public repositories, with India and the USA leading the list.

Developers who publish code for the first time do so in existing open-source projects. In addition to the still popular home automation software Home Assistant, the Zen Browser, among others, saw strong growth. Most contributors are found in AI infrastructure projects such as VLLM, Codex, Hugging Face, or Ollama.

(dzi)

Don't miss any news – follow us on Facebook, LinkedIn or Mastodon.

This article was originally published in German. It was translated with technical assistance and editorially reviewed before publication.