Developer output
3.8M
Public GitHub commits attributed to the Copilot coding agent
Copilot's coding agent acts through its own workflow, so this is exact. Its editor completions land under the developer's identity and a separate command-line surface is counted apart, so total Copilot output is far larger than this.
Last full week (ending 2026-10-18): 19,535 signed commits (-12.2% week over week), exact count.
Distinct users: 5,116 humans opened the 12,598 most recently sampled PRs, about 2.5 PRs per person.
Merge rate: 68.2% of its pull requests get merged (78,769 of 115,560 over 8 settled weeks, 2026-08-10 to 2026-10-04). The window ends about ten days back so recently-opened PRs have had time to merge.
GitHub review activity: 4.2% of its merged PRs had at least one GitHub review record before merging; the other 95.8% had none. This is a broad weekly signal and can include bot reviews; the human-review read below comes from the enriched PR sample.
In significant repos (company-owned or 10+ stars), 80.5% of its merged PRs got a human review. Across all sampled repos, the rate is 64.2%. Human review means a non-author, non-bot GitHub reviewer in the PR-lifecycle sample.
Median time to merge: 33 min. The unreviewed median is 0.2h. The human-reviewed median is 1.24h. From a 120-PR sample merged 2026-07-30 to 2026-08-28.
Test-inclusion: 31.7% of its sampled merged PRs touch a test file. This is a rough read on production readiness.
Commits per PR: 1.6 over the last 8 weeks. Both counts come from the same visible workflow, so this is a clean measure of how many commits make up one delegated task.
Flow share: 0% of the 1999 most recent public PRs sampled (0 PRs, 2 windows, last 2026-09-27). Zero marked PRs in the sampled flow despite the historical total: current run-rate is far below the accumulated stock.
Task lens: 1.9M agent-opened PRs. Copilot's coding agent opens every task as a pull request, so this is its task-level count, exact.
Snapshot 2026-06-09, confidence High, tracked since 2026-06-09, +0.4% over 7 days. Read the methodology.