Octoverse 2026: GitHub logs 34.3 billion commits as agent-assisted work surges

A record year for activity on GitHub

According to GitHub's annual Octoverse report, published on the GitHub Blog, developers pushed code to the platform 17.9 billion times over the past 12 months — more than the 2024 and 2025 totals put together. Commits across public and private repositories climbed to 34.3 billion, more than triple the level of two years ago, while merged pull requests also more than tripled, reaching 1.3 billion, an average of 3.5 million per day.

The report dates the sharpest acceleration to around June. GitHub attributes the surge to several industry shifts arriving at once: improved underlying models, better tooling built around them, and agents that evolved from tools tied to a single session into persistent systems able to track work across days, coordinate other specialized agents, and run work in the cloud rather than only on a local machine.

Agents linked to a different pace of work

Among repositories created this year, GitHub found that the average repository containing an agent configuration file recorded eight times as many commits as the average new repository without one — evidence, the company argues, that individuals and teams building with agents from day one operate at a markedly different rate.

Pull request habits changed too. As of September, developers began more than 40% of pull requests with an agent, up from 17% in March. Model choice is fragmenting as well: in September, 61% of GitHub Copilot users worked with more than one model, spreading requests across general-purpose models, reasoning and planning models, and fast, lightweight options.

Automation and open source scale with the surge

GitHub Actions ran a job 1,440 times every second — roughly 124 million runs per day, more than three times last year's total. Dependabot pull requests, used to keep dependencies current, passed 30 million merges, about double the pace of 2025.

Public contributions reached 2.5 billion in 2026, more than double last year's 1.12 billion. The report points to individual maintainers behind the aggregate: @kinlane logged over a million contributions using agentic workflows to keep a catalog of public API data current, @kim-em made more than 178,000 contributions to mathlib, a community library of computer-checked mathematics, and @andig contributed over 21,000 times to evcc, an open-source EV charging and home energy management system.

Growth came from both new and long-standing projects. From June onward, about 55% of all commits originated from newly created repositories. Yet developers who pushed code in the past 12 months have been on GitHub for an average of 36 months, more than 75% of those who opened a pull request in August pushed code again in September, and the average developer now works across 10% more repositories than a year ago.

More developers, from more places

GitHub now counts more than 225 million developers. The United States (38.3 million) and India (32 million) remain the largest communities, and each added more than 10 million developers this year; Brazil added 3.28 million, the third-largest increase. China (12.4 million) and Brazil (10.2 million) follow in total size, ahead of Japan, the United Kingdom, Indonesia, Germany, Russia and France.

For the first time, Octoverse breaks out public repositories by industry, reporting that repositories serving mining, construction, health care and agriculture have each grown several times over since 2022.

GitHub also says it is running ahead of its own projections. One year into last year's five-year forecasts — which put India at 57.5 million developers and the US at 54.7 million by 2030 — every projected country has closed at least a fifth of the distance to its 2030 target, with the US at 38.6% and India at 28.3%. The company cautions that looking further ahead is getting harder: it ran more than 40 forecasting models out to 2031, and they no longer agree on where growth goes next.

Why it matters

The Octoverse figures are among the clearest quantitative signals yet that AI agents have moved from experiments to routine infrastructure in software development. When developers start more than 40% of pull requests with an agent, and agent-configured repositories commit at eight times the average, agent use is no longer a side channel — it correlates with output at platform scale. That shift raises questions the report does not answer, such as who reviews the rising volume of machine-drafted code and how open-source maintainers absorb it. There is also a measurement caveat: the data comes from GitHub's own telemetry, and the company has a commercial stake in the agent story it tells. Even so, one detail stands on its own — GitHub's own forecasting models disagree about the next five years, meaning even the platform hosting this activity is no longer confident where the trend lands.