🔍 Read the full analysis: Are AI Agents Creating More GitHub Pull Requests Than Teams Can Review? on ThorstenMeyerAI.com
Get business pricing on tech for your team
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
GitHub’s Octoverse report says more than 40% of pull requests on its platform in September 2026 began with an AI coding agent, up from 17% in March. The company also introduced tools for AI-assisted review, but its figures do not show whether merged changes were correct or how the agent share breaks down between public and private repositories.
AI coding agents started more than 40% of pull requests on GitHub in September, up from 17% in March, according to the company’s 2026 Octoverse report. GitHub says that amounted to 110 million agent-started pull requests in one month, a surge that puts pressure on teams to review and approve code as quickly as agents can draft it.
GitHub defines an agent-started pull request as one whose first draft came from a coding agent the platform could detect. The company describes the count as a floor because it may not identify every instance of agent use. A pull request is a proposed code change submitted for review before it is incorporated into a project; the report’s measure does not mean an agent wrote every line or that the resulting change was accepted as correct.
GitHub 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 had a merge rate of about 86% in September. A merge records that a change entered a codebase, not that it was free of defects or produced a good outcome. The report also says 87% of agent-started pull requests changed more than one file, versus 79% of other pull requests.
Alongside the report, GitHub introduced Copilot code review Max, which uses reviewers from multiple model providers and combines their findings, and Copilot Fusion, a system that routes tasks to models. GitHub also made Copilot approvals eligible to count toward required approvals in some workflows. The company estimates an average Copilot review costs $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 figures point to a change in where software teams may spend time: producing a first draft is becoming faster, while reading, testing and deciding whether to accept it still require oversight. If the number of proposed changes rises faster than teams can review them, organizations may face longer queues, heavier reviewer workloads or pressure to accept changes with less scrutiny. The report does not establish that these effects are happening across GitHub, but its scale makes review capacity a practical planning issue.
GitHub’s new review products also put a price on automated assistance. Companies weighing them will need to account for per-review charges and Actions usage, as well as the staff time needed to verify findings. Pull request counts alone are a weak measure of productivity: teams would also need to track rework, reversions, defects and incidents to judge whether faster drafting is improving software delivery.
For developers, the shift may increase the value of writing precise tasks and critically examining diffs. It may also affect how new programmers learn: small, bounded changes have often served as early assignments, and those are among the tasks coding agents can take on. The report does not measure changes to hiring or training, so the longer-term effect on junior roles remains open.
Products Worth Considering
Samsung 27-inch Odyssey OLED G6 QD-OLED QHD Gaming Monitor
LG 34GX90SA-W 34-inch Ultragear WQHD OLED Curved Gaming Monitor
Samsung 27-Inch Odyssey QD-OLED G8 4K 240Hz Gaming Monitor
As an affiliate, we earn on qualifying purchases.
How GitHub Counts Agent Work
The Octoverse report is GitHub’s annual review of activity on its platform. The 2026 edition covers October 1, 2025, through September 30, 2026. GitHub says its metrics generally include both public and private repositories unless a metric is specifically excluded; the report does not provide a public-versus-private breakdown for the agent-started pull request figure.
The report uses several related labels, including “agent-started,” “agent-generated” and “agent-opened.” Its definition for the headline figure is narrower and more specific: the first draft must have come from an agent GitHub can detect. GitHub also recalculated 2024 and 2025 figures under updated definitions and reporting windows, so comparisons with earlier Octoverse reports are not direct.
The increase comes amid broader growth in platform activity. GitHub reports 1.3 billion merged pull requests during the year ending in September, compared with 538 million in 2025 and 402 million in 2024. It also says 55% of monthly active Copilot users used agent features, up from 35% a year earlier. Those are company-reported platform metrics, not an independent assessment of code quality.
“A pull request whose first draft came from a coding agent that we can detect.”
— GitHub, in its Octoverse report
What the Report Cannot Establish
The 40%-plus figure is a platform measurement, not a complete count of all AI-assisted coding. GitHub says undetected use is missing, and it does not disclose how many of the counted pull requests came from public repositories versus private company code. Its wording also distinguishes an agent’s first draft from the eventual pull request, which may have been substantially rewritten by a person.
The report does not show whether merged agent-started changes contained defects, required more follow-up work or caused production incidents. It also does not establish that the rise in agent use caused a review backlog. GitHub reports merge rates and change sizes, but those measures do not reveal review time, reviewer workload, or the quality of the final code.
It remains unclear how often organizations will allow AI-generated review findings to satisfy approval requirements in practice, and how teams will balance the stated review prices against the cost of human oversight. GitHub’s estimates do not include the separate cost of GitHub Actions minutes.
Teams Weigh Review Rules and Costs
GitHub’s newly announced review and routing products give teams options to test as agent-generated changes become more common. Organizations will need to decide which repositories or types of changes can use automated review, whether Copilot approvals can count toward required approvals, and when a human reviewer must sign off. GitHub says developers remain responsible for shipped code, while its cloud agent cannot approve or merge its own pull requests.
The next useful evidence will be operational: how review queues change, how often agent-authored changes need rework, and whether defect or incident rates move as adoption grows. GitHub’s current report does not supply those measures. Until such data is available, the September count shows the volume of detected agent-started work, but not whether teams are keeping pace with it or whether the resulting software is better.
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 that GitHub could detect. The definition does not mean an agent necessarily wrote the final version without human changes.
Did AI agents start more than 40% of all GitHub code changes?
No. GitHub’s figure refers to pull requests in September 2026, not every code change. The company says it may miss agent use it cannot detect, and it has not published a public-versus-private repository breakdown for this metric.
Does a high merge rate mean agent-written code is correct?
No. GitHub says 90% of detected agent pull requests were merged within two weeks, but merging is not a measure of correctness. The report does not provide defect or incident rates for those changes.
How much does GitHub estimate AI code review costs?
GitHub estimates an average Copilot review at $0.25 for Lite, $1.50 for Balanced and $10 for Max. GitHub Actions minutes are additional, and the stated figures are company estimates.
Source: ThorstenMeyerAI.com
