Skip to content
Main Site News Console

GitHub Octoverse 2026 Annual Report

· Simon Willison Translated
技巧LLM

GitHub Octoverse 2026 Report. I’m attending GitHub Universe today (GitHub gave me a free ticket through their “GitHub Stars” program). Jay Parikh, EVP, just released this year’s Octoverse report, highlighting the following figures:

Figure 1: Photo of the conference keynote stage: a wide curved screen displays three line graphs side by side, with a speaker dressed in a black sweater and dark jeans standing on the right beneath an overhead lighting rig. Above each chart is a large, colorful headline figure; the shaded line plots stay low and nearly flat on the left, then curve sharply upward on the right, ending with a dot at the peak. On the left in green: "220M" (monthly merged PRs), with Y-axis at 0M, 50M, 100M, 150M, 200M. In the center in magenta: "7.38B" (monthly commits), with Y-axis at 0B, 2B, 4B, 6B, showing the most dramatic, near-vertical final surge. On the right in orange: "30.3M" (monthly new repositories created), with Y-axis at 0M, 10M, 20M, 30M, exhibiting a more meandering and gradual climb. All three graphs share X-axis labels '23, '24, '25, '26.

This provides an incredibly vivid look at the impact of agentic engineering on GitHub. In 2026, monthly merged pull requests (PRs) tripled in under a year, reaching 220 million per month. Monthly commits more than tripled to 7.38 billion per month, with most of that growth occurring in the second half of the year. The rate of new repository creation doubled this year, reaching 30.3 million per month.

The report also includes new data broken down by industry, along with dedicated figures for repositories involving “interoperability or reverse engineering.” The methodology section explains how they compiled these metrics:

Top countries and regions by pushes to public repositories (calibrated for profile location). Aggregated counts of pushes per country/region across the Octoverse year. To eliminate VPN-induced bias, we used LLM classification to extract ISO 3166-1 alpha-2 codes from the profiles of GitHub users who contributed pushes during this period […]

Industry. A fine-tuned multilingual embedding model was used to classify a random sample of public repositories created each year between 2022 and 2026 into North American Industry Classification System (NAICS) sectors, using repository names, descriptions, topics, and README information. […]

Interoperability. Count of public repositories created each year whose name, description, topics, or README mention interoperability, reverse engineering, or related terms (such as clean room implementation, drop-in replacement, or protocol re-implementation).