Skip to content
back to the archive page
#OpenSource

[2/4] Comparing GitHub’s Octoverse Reports for 2023–2025 (Category OpenSource)

#OpenSource #AI #ML #Software #Architecture #Processes #DevEx #Devops #Metrics #Engineering

Let us continue this comparative review of GitHub’s reports by discussing AI :)

AI projects The story here is growth in both numbers and adoption.

  • In 2023, AI projects took off. Generative-AI project numbers rose 248% compared with 2022. By mid-2023, new generative-AI repositories had already exceeded the total for all of 2022. Such projects entered the top 10 by contributor count for the first time, including Stable Diffusion and LangChain. Open-source developers’ experimentation with OpenAI and other models moved AI from a research niche into the mainstream. The number of developers contributing to generative-AI projects grew 148% over the year.
  • In 2024, that wave continued. GitHub had around 137 000 public generative-AI projects by year end, nearly double the previous year, at 98% growth. Contributions to AI repositories rose 59%. Of the 10 fastest-growing open-source projects by contributor count, 6 concerned AI infrastructure, such as models, orchestration, or optimization. The community was increasingly building AI tools, beyond using AI to generate code.
  • By 2025, AI projects had become mainstream. GitHub had about 4.3 million AI-tagged repositories, including private ones. More than 1.1 million public repositories used LLM SDKs; 693 thousand of those had been created in the previous year, up 178% from August 2024. Nearly one in four new repositories in 2025 contained AI/ML components. And 60% of the top 10 open-source projects by contributor count had an AI focus.

From 2023 to 2025, AI moved from an experimental trend to an integral part of the development ecosystem.

Developers’ adoption of AI tools

  • In 2023, 92% of developers had tried AI coding tools such as GitHub Copilot in some form. That points to near-universal exposure to AI assistants in development workflows. Nearly a third of starred open-source projects on GitHub had at least one maintainer using Copilot. GitHub offered it free to OSS maintainers, and many took up the offer: AI code completion became a common helper in open source. Its integration into GitHub itself seems to have helped adoption :)
  • In 2024, GitHub reported that more than 1 000 000 students, teachers, and open-source developers had used Copilot through its free education and OSS program, a 100% increase over the year. GitHub kept extending Copilot’s reach and introducing newcomers to it from their first days on the platform.
  • Copilot Free’s release in late 2024 led to a visible jump in registrations and activity. In 2025, newcomers therefore started using AI almost immediately: 80% used Copilot within their first week on GitHub. For a new generation of coders, an AI assistant had become an expected default. Nearly every newcomer was learning AI code generation immediately, reinforcing its place as a standard part of development work.

#AI #ML #Software #Architecture #Processes #DevEx #Devops #Metrics #Engineering #OpenSource #Architecture