🔍 Read the full analysis: GitHub’s AI-Started Pull Requests Put Code Review At The Center on ThorstenMeyerAI.com
Get the latest gadgets delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
TL;DR
GitHub’s Octoverse report says coding agents started 110 million pull requests in September 2026, more than 40% of the platform’s total, up from 17% in March. The company also announced AI review and model-routing tools, while maintaining that developers remain accountable for code that ships.
GitHub says coding agents started 110 million pull requests on its platform in September 2026, accounting for more than 40% of the month’s total and rising from 17% in March. In its Octoverse report, released October 28 at the company’s Universe conference, GitHub paired the figures with new tools for AI-assisted code review, putting the growing work of checking and approving code in focus.
GitHub defines an agent-started pull request as one whose first draft came from a coding agent the company can detect. The measure includes public and private repositories unless otherwise specified, but GitHub did not publish a breakdown between them. The company says the count is a floor because some agent use may not be detected. A pull request can still count as agent-started even if a person later changes much of the code.
The report says the share rose from 17% in March to more than 40% in September, and that September’s 110 million figure was about six times March’s volume. GitHub also reported that 90% of detected agent-started pull requests were merged within two weeks in September, compared with 83% in February. Pull requests without a detected agent stayed at about 86%. These figures describe merging, not whether the resulting changes were correct or free of problems.
GitHub announced Copilot code review Max, which draws on several AI reviewers from different model providers and combines their review; it also made Copilot approvals count toward required approvals and released Copilot Fusion, a system that routes tasks to models. GitHub estimates an average review costs $0.25 for Lite, $1.50 for Balanced and $10 for Max, with GitHub Actions minutes charged separately.
Infographic 01 · The agent share
Four in ten pull requests now start with an agent.
GitHub’s own telemetry, March to September 2026. The first draft comes from an agent; review and merge stay with people.
Merged within two weeks
Share of pull requests merged within 14 days
Merged fast is not the same as correct. GitHub publishes no revert or defect rate.
Bigger changes to review
Share of pull requests that change more than one file
More files per change means more for a reviewer to check.
What GitHub counts: an “agent-started pull request” is one “whose first draft came from a coding agent that we can detect.” Undetected agents are missing, so GitHub calls the figure a floor. Metrics cover public and private repositories unless noted.
Source: GitHub Octoverse 2026, 28 Oct 2026 (text, chart captions, glossary and methodology). All figures are GitHub’s own.
Powered by Thorsten Meyer AI
Review Becomes the New Constraint
The report points to a shift in where software teams may spend time. If agents produce more initial code drafts, the capacity to inspect changes, catch defects and decide whether to merge them can become a limiting factor. GitHub’s product announcements reflect that pressure: they focus not only on generating code, but also on review workflows, model routing and approval rules.
The consequences reach beyond tooling budgets. GitHub’s stated review prices give teams a way to estimate some costs, but the figures exclude Actions minutes and do not establish how much time a human reviewer saves. Organizations will need to track rework, reversions and incidents, rather than treating a higher pull request count or merge rate as proof of better software.
The change also affects developers’ responsibilities and training. GitHub says developers remain accountable for what ships. As agents take on more first drafts, writing a clear task and critically reading a diff become more central skills. The report’s data alone does not show how these shifts affect job roles or the quality of code in production.
Products Worth Considering
Bose QuietComfort Wireless Noise-Cancelling Over-Ear Headphones, Black
Logitech MX Mechanical Wireless Illuminated Keyboard Tactile (Graphite)
Samsung 27-inch Odyssey OLED G6 QD-OLED QHD Gaming Monitor
As an affiliate, we earn on qualifying purchases.
How GitHub Defines Agent Work
The figures come from GitHub’s Octoverse 2026, its annual report on activity across the platform. This edition covers October 1, 2025, through September 30, 2026. GitHub recalculated 2024 and 2025 figures using updated definitions and reporting windows, so earlier Octoverse numbers are not directly comparable with the new series.
GitHub’s wording varies across the report: it uses “agent-started,” “agent-generated,” “agent-authored” and “agent-opened” in different places. Its definition points to the first draft as the criterion, so agent-started is the clearest description of what the headline figure measures. The company also reported that 87% of detected agent pull requests changed more than one file, compared with 79% of pull requests without a detected agent—an indication that reviewers may need to examine broader changes, not evidence on its own of greater risk.
GitHub reported 1.3 billion merged pull requests across the platform during the year, up from 538 million in 2025 and 402 million in 2024, using the report’s recalculated series. The overall growth provides scale for the agent figures, but the report does not attribute all of the increase in pull requests to AI agents.
“A pull request whose first draft came from a coding agent that we can detect.”
— GitHub, describing its agent-started pull request measure
What the Metrics Cannot Show
The report does not say how many agent-started pull requests came from public versus private repositories, or how often agent contributions were substantially rewritten by people before review. Because GitHub counts only agent activity it can detect, the actual share may differ from the reported figure; the company characterizes its count as a floor, not a complete census.
The merge data also leaves key quality questions unanswered. A merged change is not necessarily correct, and the report does not establish whether agent-started pull requests lead to more defects, less rework or faster delivery after release. Nor does it show whether AI reviews consistently catch issues that human reviewers would miss. The quoted prices are GitHub estimates, and the stated amounts do not include GitHub Actions minutes.
Teams Face Review Decisions
GitHub’s new review and routing features give organizations more options for deciding how AI contributes to code checks and required approvals. The company’s report does not set a deadline for adopting them or provide results showing how they perform across different teams. Customers will need to decide where an AI review may count toward an approval requirement and where a human sign-off remains necessary.
The next practical test is whether added agent capacity produces dependable software without overwhelming reviewers or budgets. Teams can compare review costs and time with outcomes such as rework, reversions and production incidents. GitHub’s annual figures establish the scale of agent-started work it detected; evidence about the quality and downstream effects of that work remains to be gathered.
Key Questions
How many GitHub pull requests did agents start in September 2026?
GitHub reported 110 million agent-started pull requests in September, representing more than 40% of the month’s total, according to its Octoverse report.
What does GitHub count as an agent-started pull request?
GitHub counts a pull request when its first draft came from a coding agent the company can detect. The measure can include changes later rewritten by a person, and it may miss agent use GitHub cannot detect.
Does the merge rate prove agent-written code is reliable?
No. GitHub reported that 90% of detected agent-started pull requests were merged within two weeks in September, but merging does not establish correctness or show what happened after release.
How much does GitHub estimate AI code review costs?
GitHub estimates an average review at $0.25 for Lite, $1.50 for Balanced and $10 for Max. GitHub Actions minutes are additional, and the report does not provide a separate estimate for those charges.
Who is responsible for code that an AI agent helps produce?
GitHub says developers remain accountable for what ships. It also says its cloud agent cannot approve or merge its own pull requests, and the person who requested the change cannot approve it.
Source: ThorstenMeyerAI.com
