More than 40% of pull requests on GitHub started with an agent’s first draft by September 2026, up from 17% in March, according to GitHub’s Octoverse 2026 report (GitHub, October 28, updated November 6). GitHub says that was 110 million pull requests in September alone, about 6x March’s volume.
The report also says developers on paid GitHub Copilot plans used 47 distinct models in September, up from 29 in May, and that no single model carried more than 25% of requests that month.
Business Pill · The Model and the Harness
An evergreen Business Engineer explainer on the difference between a model and the harness around it. It teaches the general idea only and does not discuss GitHub’s figures.
The key insight: GitHub’s Octoverse 2026 report says more than 40% of pull requests on GitHub started with an agent’s first draft by September 2026, up from 17% in March, as commits rose to 34.3 billion and paid Copilot users spread requests across 47 models.
GitHub says developers pushed code to GitHub 17.9 billion times in the 12 months to September 30, “more than in 2024 and 2025 combined”. Commits across public and private repositories rose to 34.3 billion, which GitHub says is “more than three times where they stood two years ago”. Merged pull requests reached 1.3 billion, an average of 3.5 million a day, and GitHub Actions ran 45.4 billion jobs.
GitHub says it now serves more than 225 million developers. The United States (38.3 million) and India (32 million) remain the two largest communities, and GitHub says each added more than 10 million developers this year.
GitHub’s Octoverse 2026 report gives annual totals for the 12 months to September 30: commits rose from 10.0 billion in 2024 to 34.3 billion in 2026, pushes from 5.2 billion to 17.9 billion, GitHub Actions jobs from 8.0 billion to 45.4 billion and merged pull requests from 402 million to 1.3 billion. The multiples are our arithmetic.
Where Agents Show Up
GitHub says that among repositories created this year, the average repository with an agent configuration file logged eight times as many commits as the average new repository without one. It says about 87% of agent-started pull requests change multiple files, against 79% of pull requests without detected agent involvement, and that by September 90% of agent pull requests merged within two weeks, against about 86% for the comparison group.
Among GitHub Copilot’s monthly active users, GitHub says 55% now use agentic features, up from 35% a year ago, and that Copilot cloud agent, where a developer assigns a task and an agent works on it independently, grew 230% year over year, against 85% for agent chat.
GitHub’s methodology says agent signals are “a floor, since not all agent use may be detected”, and that its model-choice figures count only requests where developers selected a model themselves, excluding auto model selection and requests made with a developer’s own API key.
How This Compares With Earlier Figures
Our November 4 piece on OpenAI’s Pierre acquisition quoted Microsoft’s earnings call, where it said agent-generated pull requests on GitHub “have increased 6X over the past six months” and that “agents now open more than a third of all PRs on the platform”. The Octoverse report gives the September share as more than 40%.
By our arithmetic, 34.3 billion commits divided by 225 million developers is about 152 commits per developer for the year; because GitHub says it serves more than 225 million developers, the true average is lower, and GitHub does not report this figure. Commits in 2026 were 3.4 times the 10.0 billion GitHub reports for 2024, pushes 3.4 times (17.9 billion against 5.2 billion) and Actions jobs 5.7 times (45.4 billion against 8.0 billion).
The Structural Read
As we read GitHub’s figures, agents moved from a minority of pull requests in March to more than 40% by September, while merge rates for agent pull requests rose to 90%, slightly above the comparison group.
As we read the model data, no single model dominates paid Copilot usage: GitHub says 47 models were used in September and none carried more than 25% of requests.
As we read the volume data, GitHub’s telemetry shows the step change came in 2026: commits were 12.9 billion in 2025 and 34.3 billion in 2026.
GitHub, Octoverse 2026
“For a growing share of work, the first move is now directing an agent, and review is how developers shape what comes back”
Three Implications
FOR DEVELOPERS GitHub says 55% of Copilot’s monthly active users now use agentic features, up from 35% a year ago.
FOR MODEL MAKERS GitHub says paid Copilot users used 47 distinct models in September, up from 29 in May, and no single model carried more than 25% of requests.
FOR GITHUB’S INFRASTRUCTURE GitHub says Actions ran 45.4 billion jobs in the past 12 months, 1,440 every second.
The analysis says: “The surface, not the model, governs autonomy.”
As we read it, GitHub’s data fits that frame: agents now start more than 40% of pull requests on GitHub, while requests in paid Copilot are spread across 47 models with none above 25%. The analysis does not discuss GitHub’s Octoverse figures.
What Is Not Established
The figures are GitHub’s own telemetry, and we could not check them independently. GitHub says its 2024 and 2025 figures were recalculated on the same definitions and may differ from earlier reports. The report does not break agent pull requests down by vendor or model. The BigGo summary that prompted this check cites a Latent Space podcast with Microsoft’s Jay Parikh; we could not find that episode and use only GitHub’s published report here. This piece makes no forecast.
This Week In The AI Supercycle
The Great Convergence of Capital, Power, and Geopolitics
This week’s issue: for two years the AI build-out was priced as a supply story, and last week as a credit story. This week it was priced as a permission story, and the answer arrived as a cost, not a ban.
GitHub’s report puts agents at the start of more than 40% of pull requests by September, up from 17% in March, while commits more than tripled in two years. On the model side, paid Copilot users spread their requests across 47 models in September, and GitHub says no single model carried more than 25% of them.
A note on sourcing. Figures and quotations are from GitHub’s Octoverse 2026 report (October 28, 2026, updated November 6), based on GitHub’s own telemetry. Earlier Microsoft figures are from its earnings call as quoted in our November 4 piece. The ratios and multiples are our arithmetic. This piece is not investment advice.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
