🔍 Read the full analysis: Four In Ten GitHub Pull Requests Start With AI Agents, Shifting Review 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 in September began with a detectable coding agent, up from 17% in March. The figures measure agent-started drafts, not code quality; GitHub says 90% were merged within two weeks, while adding AI review tools and keeping final responsibility with developers.
More than 40% of pull requests on GitHub in September began with a first draft from a detectable coding agent, up from 17% in March, according to GitHub’s 2026 Octoverse report, released October 28. The company says that amounted to 110 million agent-started pull requests in September, a shift that is prompting new attention to how quickly people can review and approve code.
GitHub defines an agent-started pull request as one whose first draft came from a coding agent it can detect. The company says its count is a floor, because some agent use may not be detected. A pull request can still count as agent-started if a person later changes much of the draft. GitHub’s methodology includes public and private repositories unless it says otherwise; the report does not give a separate breakdown for those repository types.
GitHub reported that 90% of detected agent pull requests were merged within two weeks in September, compared with 83% in February. Pull requests without a detected agent had a merge rate of about 86%. Those figures show that more agent-started proposals are being accepted, but a merge is not evidence by itself that a change is correct, safe or free of later rework.
At its Universe conference, GitHub announced Copilot code review Max, which combines reviews from several AI models, and Copilot Fusion, a system that routes different steps to models. The company also said Copilot review approvals can count toward required approvals. GitHub estimates a Copilot review costs an average of $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 change in where software work may slow down. If agents produce more first drafts, the time required to read, test and assess each proposed change can become a limiting factor, even when code generation itself is fast. GitHub’s new review products address that pressure, but their stated prices also make review a metered expense that teams may need to budget alongside compute and development tools.
The distinction between approval volume and software quality matters for companies deciding whether to rely on AI review. A high merge rate does not report how often a change required later fixes, was reverted or caused an incident. Teams will need outcome measures as well as counts of drafts and approvals to judge whether agent use is helping.
For developers, more agent-written code can shift the work toward writing precise instructions and critically evaluating diffs. It may also affect how newer engineers gain experience: routine starter tasks are among the kinds of work agents can take on. The report does not measure that effect, but the change in the mix of work could make mentoring and deliberate practice more important.
Products Worth Considering
Bose QuietComfort Wireless Bluetooth Headphones with Active Noise Cancellation, Mic, USB‑C Charging, Deep Bass, 24‑Hour Battery – White Smok
Sony WH-1000XM6 Wireless Noise-Canceling Over-Ear Headphones, Blue, Bundle
Logitech MX Mechanical Wireless Illuminated Keyboard Tactile (Graphite)
As an affiliate, we earn on qualifying purchases.
How GitHub Defines Agent Work
The Octoverse is GitHub’s annual report on activity across its platform. This edition covers October 1, 2025, through September 30, 2026. GitHub says its historical 2024 and 2025 figures were recalculated using updated definitions and the report’s measurement window, so comparisons with older editions may not be direct.
The report says merged pull requests rose to 1.3 billion in the 12 months to September, from 538 million in 2025 and 402 million in 2024. It also reports 17.9 billion code pushes and 34.3 billion commits during the period. These platform-wide totals provide scale for the rise in agent activity, but they do not isolate how much overall growth agents caused.
GitHub says 55% of monthly active Copilot users now use agent features, up from 35% a year earlier, and use of its cloud agent grew 230% year over year. The report also says agent-started pull requests more often changed multiple files: 87%, compared with 79% of pull requests without a detected agent. That can increase the amount of related code a reviewer must understand, though file count alone does not establish complexity or risk.
“A pull request whose first draft came from a coding agent that we can detect.”
— GitHub, in its Octoverse methodology
What the Metrics Leave Out
The headline figure is limited to agents GitHub can detect, and the report does not publish the share of agent-started pull requests by public versus private repository. GitHub’s wording also varies in places, using terms such as “agent-generated,” “agent-authored” and “agent-opened”; its definition identifies the key measure as a pull request started from an agent’s first draft.
The report does not establish whether merged agent-started changes perform as well as human-started changes over time. It provides no comparable data here on rework, reversions, production incidents or code quality. Nor does the reported merge rate show how much human review each pull request received or whether AI review changed the outcome.
GitHub’s cost figures are estimates for average reviews, not a complete estimate of company spending. The total cost will depend on review volume and GitHub Actions usage, among other factors. It is also not yet clear how organizations will set policies for AI-generated approvals or measure their effect on review workloads.
Teams Set Review and Approval Rules
GitHub has released the review and routing features described in its conference announcements, while developers and organizations will need to decide how to use them. Companies can set policies for where AI approvals count, determine which changes require a human reviewer, and account for review credits and Actions minutes in budgets.
For teams evaluating whether agent use improves delivery, the next useful evidence will go beyond the number of pull requests opened or merged. Tracking review time, follow-up fixes, reversions and incidents could help distinguish faster drafting from better outcomes. GitHub’s report does not provide those measures for the agent-started changes, so their impact remains an open question.
GitHub’s annual report gives a platform-wide snapshot through September 2026. Future reports or company-level data may show whether the growing share of agent-started work changes review effort and software quality, but no specific next measurement date or outcome has been provided.
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. People may revise that draft afterward and it still counts. GitHub says its total may miss agent use it cannot detect.
How many pull requests did agents start?
GitHub says detectable agents started 110 million pull requests in September 2026, representing more than 40% of the platform’s pull requests that month. The share was 17% in March, according to the report.
Does the merge rate show that agent-written code is good?
No. GitHub reported that 90% of detected agent pull requests were merged within two weeks in September, but merging does not by itself establish quality. The report does not provide comparable data on later fixes, reversions or incidents.
How much does GitHub’s AI code review cost?
GitHub estimates average review costs of $0.25 for Lite, $1.50 for Balanced and $10 for Max. GitHub Actions minutes are additional, and actual spending will depend on usage.
Who is responsible for code written by GitHub’s agent?
GitHub says developers remain accountable for code that ships. It says its cloud agent cannot approve or merge its own pull request, and the person who requested a change cannot approve it.
Source: ThorstenMeyerAI.com
