🔍 Read the full analysis: GitHub Pull Requests Are Increasingly AI-Started; Review Is Shifting Too 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 2026 Octoverse report says more than 40% of pull requests on its platform in September began with a coding agent’s first draft, up from 17% in March. The company is adding AI review and model-routing tools, but says developers remain accountable for shipped code; its figures measure detected agent activity, and a merge does not establish that code is correct.
More than 40% of GitHub pull requests in September started with a coding agent’s first draft, up from 17% in March, according to the platform’s 2026 Octoverse report. The growth is prompting GitHub to focus new product releases on review, approval and model routing—the steps that determine whether AI-written code is ready to merge.
GitHub says the September figure represents 110 million agent-started pull requests, about six times the March volume. It defines an agent-started pull request as one whose first draft came from a coding agent the company can detect. GitHub describes its count as a floor because some agent use may not be detected. The report’s methodology includes public and private repositories unless specified otherwise, but GitHub did not provide a public-versus-private breakdown for this metric.
In September, 90% of detected agent pull requests were merged within two weeks, compared with 83% in February. Pull requests without a detected agent remained at about 86%, GitHub reported. Those figures track merging, not whether changes were correct, secure or free of later rework. The report also says 87% of agent-started pull requests changed more than one file, compared with 79% of other pull requests.
At its October 28 Universe conference, GitHub announced Copilot code review Max, which uses several AI reviewers from different model providers and combines their feedback. It also said Copilot review approvals can count toward required approvals and introduced Copilot Fusion, a system that routes individual tasks to models. GitHub’s estimated review prices are $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 figures suggest that generating a proposed change is becoming faster than evaluating it. When agents produce more pull requests—and often touch multiple files—teams may face additional demands on reviewers’ time, even if many proposals are merged quickly. Review capacity and quality controls could shape how much value organizations get from AI coding tools.
GitHub’s new review products turn that pressure into a product and budget question: teams can choose different review levels, but each has a stated price and may also use Actions minutes. Organizations will need to decide where AI feedback is useful, whether an AI approval can meet their rules, and how to track outcomes beyond the number of pull requests merged. Reverts, incidents and rework would provide a different picture from merge rates alone.
For developers, the shift may change the work that takes the most attention. Clear task descriptions and careful reading of diffs become more consequential when an agent supplies the first draft. The source material also raises a workforce concern: routine starter tasks that once helped junior developers gain experience may be among the tasks agents take on. The report does not measure how that affects hiring or training.
Products Worth Considering
Logitech MX Mechanical Wireless Illuminated Keyboard Tactile (Graphite)
Samsung 27-Inch Odyssey QD-OLED G8 4K 240Hz Gaming Monitor
LG 34GX90SA-W 34-inch Ultragear WQHD OLED Curved Gaming Monitor
As an affiliate, we earn on qualifying purchases.
How GitHub Defined Agent-Started
The Octoverse is GitHub’s annual report on activity across its platform. This edition covers October 1, 2025, through September 30, 2026. GitHub recalculated its 2024 and 2025 figures using updated definitions and a revised reporting window, so earlier Octoverse totals are not directly comparable without that adjustment.
The report uses several related labels—including “agent-started,” “agent-generated” and “agent-authored”—but its definition centers on the first draft coming from a detectable coding agent. A pull request can still count if a person substantially edits it afterward. GitHub’s reported share is therefore a measure of detected starts, not a measure of how much code remains AI-written in the final merged change.
The platform’s broader activity also grew during the reporting year: GitHub reported 1.3 billion merged pull requests, compared with 538 million in 2025 and 402 million in 2024. That growth provides scale for the agent figures, but the report does not establish how much of the overall increase was caused by coding agents.
“A pull request whose first draft came from a coding agent that we can detect.”
— GitHub, in its Octoverse 2026 methodology
What the Pull Request Figures Miss
The report does not disclose how many agent-started pull requests came from public repositories versus private company code, nor does it quantify the rate of undetected agent use. Because the definition is based on the first draft, the percentage also does not show how much of each final change was written by an agent after human edits.
GitHub’s merge statistics do not establish code quality or the frequency of defects, security issues, reverts or later fixes. The reported 90% merge rate also uses a two-week window and applies to detected agent pull requests in September; it should not be read as proof that those changes performed well in production. The report gives no comparative outcome data for the new AI review tools.
It is also unclear how teams will apply Copilot approvals in practice, how much review demand the new products will reduce, and what total costs will be once Actions minutes and organizational review requirements are included. GitHub’s prices are estimates for review tiers, not evidence of savings.
Teams Will Test Review Policies
The immediate next step is likely to be adoption and evaluation of GitHub’s review features as teams decide whether AI-generated feedback should supplement human review or count toward approval requirements. The company’s announced offerings provide different review levels and model routing, but the material does not give a timetable for broader availability or independent performance results.
Organizations using coding agents will need to set approval rules and monitor rework, reversions and production incidents alongside merge rates. GitHub’s next reporting and product updates may clarify how widely the tools are used and whether AI review changes those outcomes. For now, the firmest measure is the rise in detected agent-started pull requests—not evidence that the review bottleneck has been solved.
Key Questions
What does GitHub mean by an agent-started pull request?
GitHub counts a pull request when its first draft came from a coding agent it can detect. The change may be edited by a person afterward, and some agent activity may go undetected.
How many pull requests did agents start in September?
GitHub reported 110 million, representing more than 40% of pull requests on the platform that month. It said the figure was about six times the March volume.
Does the 90% merge rate mean agent-written code was correct?
No. GitHub said 90% of detected agent pull requests were merged within two weeks, but merging is not a quality or correctness measure. The report does not provide corresponding defect or rework rates.
How much does GitHub estimate AI code review costs?
GitHub estimates $0.25 for Lite, $1.50 for Balanced and $10 for Max per review, plus GitHub Actions minutes. These are stated estimates, not a full accounting of organizational costs.
Who is responsible for code that an agent proposes?
GitHub says developers remain accountable for what ships. It says its cloud agent cannot approve or merge its own pull requests, and the person who requested a change cannot approve it either.
Source: ThorstenMeyerAI.com
