🔍 Read the full analysis: GitHub PRs And AI Agents: A Changing Split Between Starting And Reviewing on ThorstenMeyerAI.com
Get business pricing on office and shipping supplies
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
GitHub’s Octoverse report says coding agents started more than 40% of pull requests on the platform in September 2026, up from 17% in March. At the same time, GitHub announced AI review and routing tools, putting more focus on checking and approving code; the company says human developers remain accountable for what ships.
Coding agents started more than 40% of pull requests on GitHub in September, up from 17% in March, according to the company’s Octoverse report released October 28. GitHub also introduced AI review and model-routing tools, reflecting a shift in the work surrounding code: as agents draft more changes, developers and teams must check them before they are merged.
GitHub said the September figure represents 110 million agent-started pull requests, about six times the volume it recorded in March. The company defines an agent-started pull request as one whose first draft came from a coding agent it can detect. It describes the count as a floor because some agent use may not be detected. A pull request is a proposed code change submitted for review before it is incorporated into a project.
GitHub reported that 90% of detected agent pull requests were merged within two weeks in September, compared with 83% in February. The rate for pull requests without a detected agent was about 86%. Those figures measure whether changes were merged, not whether they were correct, safe or free of later rework.
Among the tools announced this week, Copilot code review Max combines reviews from several AI models into one result, while Copilot Fusion routes different steps to models. GitHub also said Copilot review approvals can count toward required approvals. The company estimates an average review costs $0.25 in Lite mode, $1.50 in Balanced and $10 in 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 producing an initial code proposal is becoming less of a bottleneck for some GitHub projects, while human review capacity and approval practices take on greater importance. If agents generate changes quickly, teams still need time and expertise to assess whether those changes fit the codebase and meet quality and security requirements.
That can affect budgets as well as workflow. GitHub’s estimated per-review prices make automated review a metered cost, separate from any additional Actions usage. Companies will need to decide where AI review is useful, when it may count toward approval requirements, and how to measure results beyond the number of pull requests or merge rates. Rework, reversions and incidents provide different information from throughput alone.
For developers, the change may put more weight on writing clear task instructions and reading diffs critically. The source report also raises a workforce concern: small coding tasks can serve as entry points for junior developers, and those are among the tasks agents can take on. How that affects training and hiring is not established by the figures.
Popular right now
Bose QuietComfort Wireless Noise-Cancelling Over-Ear Headphones, Black
Bose QuietComfort True Wireless Earbuds with Active Noise Cancellation, Black
De'Longhi Magnifica Evo Automatic Espresso Machine with Milk Frother – Silver
As an affiliate, we earn on qualifying purchases.
How GitHub Counts Agent Work
The Octoverse report covers activity from October 1, 2025, through September 30, 2026. GitHub’s methodology says metrics include public and private repositories unless otherwise stated; the agent share is not marked as an exception. The company did not provide a public-versus-private breakdown for this measure.
The wording matters: GitHub’s definition counts a pull request when its first draft came from a detectable coding agent. A person may revise much of the draft and the pull request still qualifies. The report uses nearby terms such as “agent-generated” and “agent-authored” in charts, but “agent-started” most closely matches the stated definition. GitHub also recalculated 2024 and 2025 figures using updated definitions and time windows, so older Octoverse totals are not directly comparable.
The rise comes amid broader growth in platform activity. GitHub reported 1.3 billion merged pull requests in the year to September, compared with 538 million in 2025 and 402 million in 2024. It said 55% of monthly active Copilot users now use agent features, up from 35% a year earlier. These platform-wide totals provide scale but do not show how much of the overall growth was caused by agents.
“A pull request whose first draft came from a coding agent that we can detect.”
— GitHub’s Octoverse report
Limits of the September Measure
The reported share is subject to incomplete detection, and GitHub has not disclosed how many agent-started pull requests it may be missing. The company also has not supplied a split between public and private repositories, or a breakdown showing how much human revision typically follows an agent’s first draft.
Merge rates do not establish code quality or downstream outcomes. The report does not show whether agent-started changes led to more defects, rework or security problems, nor does it provide a direct comparison that accounts for the different sizes and types of work in each group. GitHub said agent pull requests more often changed multiple files—87%, compared with 79% for pull requests without a detected agent—but that measure alone does not determine review difficulty or risk.
The precise effect of AI approvals is also unsettled. GitHub says its cloud agent cannot approve or merge its own pull requests, and the person who requested a change cannot approve it either. The announced ability for Copilot reviews to count toward required approvals raises questions about how organizations will configure their rules and what human checks they will retain. The material provided does not specify adoption, customer outcomes or the full terms that apply to every review mode.
Teams Set Review Rules
The immediate next step is for teams using GitHub’s new review features to decide how they fit into existing code-review policies. That includes setting limits on when AI-generated approvals count, assigning human responsibility, and tracking review spending alongside Actions usage. GitHub’s release and pricing estimates give organizations a basis for testing the tools, but do not establish how much time or money they will save.
Further reporting will be needed to assess whether the increase in agent-started pull requests produces lasting changes in delivery speed or software quality. Useful evidence would include comparable data on rework, reverted changes, defects and incidents, as well as clearer measurement of agent detection and human edits after the first draft. Until then, the confirmed development is higher measured agent activity and a new set of review tools—not proof that merged agent code is correct or that AI review can replace human judgment.
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 the company can detect. Human edits made afterward do not necessarily change that classification.
How many pull requests did agents start in September?
GitHub reported 110 million agent-started pull requests in September 2026, representing more than 40% of the platform’s pull requests that month, according to its report.
Does a high merge rate mean agent code is correct?
No. GitHub’s figures describe whether pull requests were merged within two weeks. A merge is not a measure of correctness, and the report does not establish downstream quality or incident rates.
What does GitHub estimate AI code reviews cost?
GitHub estimates an average Copilot review at $0.25 for Lite, $1.50 for Balanced and $10 for Max. GitHub Actions minutes are additional, according to the source material.
Who is accountable for code written by an agent?
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 a change cannot approve it.
Source: ThorstenMeyerAI.com
