🔍 Read the full analysis: How To Keep Up With GitHub Reviews As AI Agents Open More Pull Requests on ThorstenMeyerAI.com
Get business pricing on monitors, keyboards and dev gear
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
GitHub’s Octoverse 2026 report says more than 40% of pull requests in September began with a detectable coding agent, up from 17% in March. GitHub introduced tools aimed at reviewing and routing this growing volume, while keeping human responsibility for shipped code. Its review tools have estimated per-review costs, and the report’s merge figures do not establish code quality.
More than 40% of pull requests on GitHub in September began with a coding agent’s first draft, up from 17% in March, according to GitHub’s Octoverse 2026 report. The company released the figures at its October 28 Universe conference, alongside new tools for reviewing and routing AI-generated code as software teams face a growing volume of changes to assess.
GitHub says the September figure represents 110 million agent-started pull requests in one month, roughly six times the March volume. The company defines these as pull requests whose first draft came from a coding agent it can detect. It describes the count as a floor because some agent use may go undetected; the report does not give a separate share for public and private repositories.
The report says 90% of detected agent pull requests were merged within two weeks in September, compared with 83% in February. Pull requests without a detected agent held at about 86%. A merge records acceptance into a codebase; it does not show that a change was correct, safe or free of later rework.
GitHub announced Copilot code review Max, which uses reviewers from multiple model providers and combines their findings, and Copilot Fusion, which routes tasks to models. It also said Copilot review approvals can count toward required approvals. GitHub estimates reviews cost $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 Capacity Becomes the Constraint
The report points to a shift in where software work can get stuck: producing a first draft is faster, while checking a larger number of proposed changes still takes reviewer attention. GitHub says 87% of agent-started pull requests changed more than one file, compared with 79% of those without a detected agent. That may add to the work of understanding how a change fits together, though the figures do not measure review time or defect rates.
For engineering teams, the immediate operational question is how to allocate human review time, AI review credits and compute costs. Teams will need to decide when an AI review can count toward required approvals and how to check whether faster merging leads to more rework, reversions or incidents. Pull-request volume alone cannot answer whether the process is producing better software.
For developers, the change may put greater weight on writing clear task instructions and critically reading diffs. Routine, small tasks can also be learning opportunities for early-career engineers; if agents take on more of them, teams may need other ways to teach code review and judgment.
Popular right now
Ninja CREAMi Scoop & Swirl 13-in-1 Ice Cream and Soft Serve Maker, NC701
Ninja DualZone Air Fryer 10 Qt Smart 6-in-1 Grey (DZ950)
Ninja Hot & Iced XL Coffee Maker with Rapid Cold Brew, Black, CM371
As an affiliate, we earn on qualifying purchases.
How GitHub Defines Agent Work
GitHub’s Octoverse report covers October 1, 2025, through September 30, 2026, and describes trends on its own platform. Its agent measure concerns how a pull request started, not how much of the final code an agent wrote. A human may substantially change an agent’s first draft and the pull request still counts as agent-started.
The report’s wording varies between “agent-started,” “agent-generated,” “agent-authored” and “agent-opened.” GitHub’s definition—a first draft from a detectable coding agent—is the most precise basis for describing the headline statistic. The company also recalculated its 2024 and 2025 figures using new definitions and a new reporting window, so earlier Octoverse totals should not be compared without accounting for that change.
The broader platform saw substantial growth too: GitHub reported 1.3 billion merged pull requests during the year, up from 538 million in 2025, and 34.3 billion commits. Those totals show the scale of activity but do not isolate how much growth came from agents. GitHub also said 55% of monthly active Copilot users used agent features, compared with 35% a year earlier.
“A pull request whose first draft came from a coding agent that we can detect.”
— GitHub, describing the report’s methodology
What the Figures Cannot Show
The report does not disclose how many agent-started pull requests came from public versus private repositories, nor how many agent uses its detection misses. It also does not establish whether the higher merge rate reflects better code, different review practices or other factors. A merged change is not, by itself, evidence of correctness.
GitHub’s cost figures are estimates for the named review modes; actual bills may also depend on usage and GitHub Actions minutes. The supplied report details do not specify how organizations should configure AI approvals in every setting or provide independent results comparing AI review with human review.
How Teams Will Measure Review
GitHub has released the review and routing features alongside its report, but the material provided does not give a rollout schedule for every capability or a date for the next major measurement update. Engineering teams adopting them will need to track review time, corrections, reverted changes and incidents alongside the number of pull requests agents open or that reviewers approve.
The next useful evidence will be whether these systems help teams keep review quality steady as agent activity grows, and at what cost. GitHub’s figures establish a rapid rise in detected agent-started work; they do not yet show that review bottlenecks have been resolved.
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. Later human editing does not change that classification.
Does a high merge rate mean agent code is reliable?
No. GitHub reports that 90% of detected agent pull requests were merged within two weeks in September, but merging does not establish correctness or code quality.
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. The figures are estimates, not a guarantee of an organization’s total bill.
Can a GitHub coding agent approve its own pull request?
GitHub says its cloud agent cannot approve or merge its own pull requests. It also says the person who requested the change cannot approve that pull request.
Source: ThorstenMeyerAI.com
