Octoverse: Commits, pushes, reach new highs in 2026 as developers build with agents
Source
github.blog
Author
GitHub Staff
Date
Why it matters
GitHub's own data shows agent use correlating with sharply higher commit and PR volume, giving engineers hard numbers on how agentic workflows are changing software output.
Key takeaways · AI-distilled
GitHub's Octoverse counts 17.9 billion pushes in the year to September 2026, more than 2024 and 2025 combined, 34.3 billion commits, over three times two years earlier, and 1.3 billion merged pull requests. Activity surged around June as agents became persistent.
By September more than 40% of pull requests started from an AI agentAn AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.Full definition →'s first draft, up from 17% in March. Agent pull requests merged within two weeks 90% of the time by September, versus about 86% for those without detected agent involvement.
Model choice is fragmenting: 61% of Copilot users who picked models used more than one, paid users touched 47 distinct models in September, and no model carried over 25% of requests. Developer-selected requests split 38 general purpose, 37 deep reasoning, 25 fast.
Among 2026 repos with an agent signal, TypeScript was the primary language 35% of the time versus 24% without one. GitHub reads this as agents favoring languages whose output is easier to check when developers leave the choice open.
Industry classification of public repos shows AI interest and agent use diverging: over 40% of repos in health care, mining and agriculture mention AI but under 3% have agent-opened pull requests, while construction leads at about 9%.
Terms in this piece · Glossary
AI agent — An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.
