🔍 Read the full analysis: What The 4-In-10 Figure Says About AI Agents And GitHub 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 says more than 40% of pull requests on its platform in September 2026 began with a detectable coding agent, up from 17% in March. The company also reported that 90% of those agent-started requests were merged within two weeks, but a merge does not establish that code was correct or safe. GitHub’s new review products and pricing put scrutiny, approval and human accountability at the center of the change.
More than 40% of GitHub pull requests in September 2026 started with a detectable coding agent, up from 17% in March, according to GitHub’s Octoverse report released October 28. The company says that amounted to 110 million agent-started pull requests in one month, a shift that raises the importance of reviewing and approving code written initially by software agents.
GitHub defines an agent-started pull request as one whose first draft came from a coding agent the company can detect. It describes the count as a floor because some agent use may not be detected. The figure includes public and private repositories; GitHub did not publish a separate breakdown for open-source and company code. A request still counts as agent-started even if a person later edits the draft.
The report says 90% of agent-started 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%, according to the report. These figures describe whether changes were merged, not whether they were correct, secure, or free of later rework.
GitHub introduced Copilot code review Max, which uses reviewers from several model providers before producing a combined review, and Copilot Fusion, a system that routes work to models for different steps. It also said Copilot review approvals can count toward repository-required approvals. GitHub estimates a review costs $0.25 for Lite, $1.50 for Balanced and $10 for Max, plus GitHub Actions minutes.
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 change is becoming faster for many GitHub users, while the work of reading, testing and approving changes remains. A larger flow of proposed code can increase demands on maintainers and engineering teams, especially when a change touches several files: GitHub says 87% of agent-started pull requests changed more than one file, versus 79% of those without a detected agent.
That makes the quality of review processes, not just the volume of generated code, a practical concern for companies. Teams may need to account for review credits and reviewer time, set rules for when an AI review can count toward approval, and monitor outcomes such as rework, reversions and incidents. Pull-request counts alone do not show whether teams are shipping better software or simply processing more proposals.
For developers, the change may affect both daily work and early-career learning. Writing a clear task for an agent and critically checking its proposed changes become more central skills. Routine starter tasks that once gave junior developers practice may be among the tasks agents can draft first; the report does not quantify what that means for hiring or training.
Popular right now
Shark Navigator Robot Vacuum with 30-Day Self-Emptying Base, White (AV2110S)
Dyson Cyclone V10 Origin Cordless Vacuum
eufy E25 Robot Vacuum and Mop Combo with HydroJet Roller Mop
As an affiliate, we earn on qualifying purchases.
How GitHub Defines Agent-Started
The Octoverse report is GitHub’s annual review of activity on its platform. Its 2026 edition covers October 1, 2025, through September 30, 2026. The report uses several related labels, including “agent-generated” and “agent-authored,” but its definition for the headline measure is specifically about a pull request’s first draft. “Agent-started” is the most precise description of what the count measures.
GitHub says it recalculated 2024 and 2025 figures using updated definitions and the new reporting window, so older Octoverse reports are not directly comparable. The platform also saw broad growth: it reported 1.3 billion merged pull requests in 2026, compared with 538 million in 2025. Those totals provide scale, but they do not establish that agents alone caused the overall increase.
GitHub reported that 55% of Copilot monthly active users used agent features, up from 35% a year earlier, and that use of its Copilot cloud agent grew 230% year over year. These are company-reported platform measures. They describe use of GitHub products and should not be read as independent measures of productivity or software quality.
“A pull request whose first draft came from a coding agent that we can detect.”
— GitHub, in its Octoverse methodology
What the 40% Figure Cannot Show
The headline does not reveal the share of pull requests that involved agents but went undetected, nor does GitHub provide a public-versus-private split for the measure. Because it counts the origin of a first draft, it also does not describe how much a human changed before review or how much of the final code was written by an agent.
The report’s merge rates do not answer whether merged agent-started changes led to bugs, security problems, reverts or less maintenance work. GitHub has not supplied those outcome comparisons in the source material. Nor is it clear from the figures how the new AI review options perform against human-only review across different repositories and kinds of code.
Finally, the stated review prices are GitHub estimates and do not include the listed Actions-minute charges. Actual spending will depend on usage and workflow choices. The impact on developer roles and junior training is also not measured by the report.
How Teams Put Review Rules in Place
GitHub has released its review offerings and model-routing product, while the report supplies a snapshot through September 30, 2026. The next practical step for organizations is deciding where agent-generated drafts and AI review fit into existing approval policies, including whether a Copilot approval can satisfy a required review.
Teams evaluating the tools will need to compare review costs with their own results: time spent checking changes, follow-up fixes, reversions and incidents. Future reporting or company-level data may clarify whether higher merge rates translate into reliable software, but the current Octoverse figures do not settle that question.
Key Questions
What does GitHub mean by an agent-started pull request?
It means the pull request’s first draft came from a coding agent GitHub could detect. A person may edit the draft later and it will still count under that definition.
Does the 40% figure mean AI wrote 40% of GitHub’s code?
No. It refers to the share of pull requests that started with a detectable agent in September, not the share of all code written by AI or the fraction of each final change authored by an agent.
Were agent-started pull requests more likely to be merged?
GitHub reported that 90% were merged within two weeks in September, compared with about 86% of requests without a detected agent. A merge does not show whether a change was correct or whether it caused later problems.
How much does GitHub estimate an 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, and actual costs depend on use.
Who is responsible for code an agent proposes?
GitHub says developers remain accountable for what ships. Its cloud agent cannot approve or merge its own pull request, and the person who requested the change cannot approve it either.
Source: ThorstenMeyerAI.com
