GitHub Octoverse 2026: merged PRs triple to 220M a month

GitHub announced its 2026 Octoverse report at GitHub Universe on 28 October 2026. Simon Willison, who was at the event on a free ticket from GitHub's "GitHub Stars" program, wrote up the announcement in his link blog. He says EVP Jay Parikh announced the report and highlighted three numbers in particular.

Merged pull requests in 2026 tripled in less than a year, to 220 million per month. Commits more than tripled, to 7.38 billion per month, with most of that increase coming in the second half of the year. The rate of new repositories doubled this year, to 30.3 million per month.

Willison reads these figures as "a pretty stark representation of the impact of agentic engineering on GitHub". That framing is his own, not Parikh's.

The report also adds new numbers by industry, and it breaks out figures for repositories that involve "interoperability or reverse engineering". Willison quotes the methodology section to show how each was built.

For the country ranking, the report counts pushes to public repositories. Pushes from each country were summed across the Octoverse year. To adjust for bias from VPN usage, an LLM classifier extracted the ISO 3166-1 alpha-2 country code from the profiles of GitHub users who contributed pushes in that period.

For industry, public repositories created from 2022 to 2026 were sorted into NAICS sectors by a fine-tuned multilingual embedding model. The model used each repository's name, description, topics and README, working from random samples per year.

For interoperability, the report counts public repositories created each year whose name, description, topics or README mention interoperability or reverse engineering. It also counts related terms such as clean room implementation, drop-in replacement or protocol re-implementation.

Key facts

  • Merged pull requests on GitHub tripled in less than a year to 220 million per month in 2026.
  • Commits more than tripled to 7.38 billion per month, with most of the increase in the second half of 2026.
  • The rate of new repositories doubled this year, to 30.3 million per month.
  • Willison attributes the surge to agentic engineering; that is his reading, not a claim attributed to GitHub's Jay Parikh.
  • The report adds industry figures (NAICS sectors, sorted by an embedding model) and a breakout of repositories mentioning interoperability or reverse engineering.

Why it matters

The three headline figures describe a sharp jump in activity on the largest code-hosting platform within a single year: merged pull requests tripled, commits more than tripled, and new repositories doubled. Willison calls it a stark representation of the impact of agentic engineering on GitHub. The commit growth is concentrated in the second half of 2026, so the change was still accelerating late in the year.

Who it affects

Anyone who builds on or depends on GitHub is the natural audience: maintainers, teams reviewing pull requests, and tool vendors. The source itself does not spell out specific groups beyond the platform. The new industry and interoperability breakouts also speak to readers who follow how software work spreads across sectors.

How to use it

Treat the three monthly figures as a reference point for how fast activity on GitHub grew in 2026: 220 million merged pull requests, 7.38 billion commits and 30.3 million new repositories per month. The methodology notes show how the country, industry and interoperability numbers were built, which matters before citing them. The full report is the place to find the industry and interoperability figures themselves.

How solid is it

The numbers come from GitHub's own Octoverse report, announced by Jay Parikh at GitHub Universe, and are relayed by Willison in a short link blog post. The methodology passages are quoted from the report's methodology section, with parts elided. The source does not define 'agentic engineering' or give any causal evidence that agents drove the growth; the attribution is Willison's own reading. The source does not say whether the monthly figures are public-only or include private repositories.

Risks and caveats

Only three figures are given, with no baseline values for the prior year, so the size of the starting point is unclear from this text. No industry breakdown, country ranking or interoperability counts appear in the post. The country ranking relies on LLM classification of profile locations and the industry split on an embedding model's sorting of sampled repositories, so both are model-assigned labels. Willison also discloses that GitHub gave him a free ticket to the event. No reaction from GitHub, Microsoft or third parties is given.

“This is a pretty stark representation of the impact of agentic engineering on GitHub.”

— Simon Willison, link blog