Enterprise AI in Numbers · Outcomes
Share of AI agent pull requests merged, versus all pull requests
84.8%
84.8% as of August 31, 2026, according to GitHub search API, our monthly counts, up from 83.3% on July 31, 2026. Of the pull requests cloud coding agents opened in the month, the share merged by the end of the following month. Parts compare agent pull requests with all pull requests.
What it means for leaders
Merge rates show how much agent work survives human review. A rate close to the baseline means agent output is treated like any contributor's; a much lower one means reviewers are spending time on work they reject.
What would change the reading: The gap between agent and all pull requests closing or widening.
The trend
The breakdown
Every reading
| Date | Value | Change | Source |
|---|---|---|---|
| 84.8% | +1.5 points | Source | |
| 83.3% | +2.4 points | Source | |
| 80.9% | −2.6 points | Source | |
| 83.5% | +0.1 points | Source | |
| 83.4% | +3.6 points | Source | |
| 79.8% | +0.4 points | Source | |
| 79.4% | +4.3 points | Source | |
| 75.1% | −2.0 points | Source | |
| 77.1% | −0.7 points | Source | |
| 77.8% | −3.0 points | Source | |
| 80.8% | +1.1 points | Source | |
| 79.7% | −3.1 points | Source | |
| 82.8% | −0.9 points | Source | |
| 83.7% | −1.7 points | Source | |
| 85.4% | — | Source |
Definition and method
- What it measures
- Of the pull requests cloud coding agents opened in the month, the share merged by the end of the following month. Parts compare agent pull requests with all pull requests.
- Population
- Pull requests opened in public GitHub repositories
- Updated
- Monthly, one month behind
- Source
- GitHub search API, our monthly counts · CC BY 4.0
Caveats
- Public repositories only.
- Agents are identified by the app account that opens the pull request, or for OpenAI Codex by its "codex/" branches, which some people also use by hand; agent work submitted through a person's own account (most local coding assistants) is not counted.
- Merged isn't the same as correct: some teams merge agent work and fix it later.
Cite this
Enterprise AI in Numbers, "Share of AI agent pull requests merged, versus all pull requests", https://lucianoayres.com/ai-brief/data/agent-pr-merge-rate/, retrieved October 25, 2026. Data: GitHub search API, our monthly counts.