23.2%

of 13,360 commits were committed by an AI agent — a floor, not an estimate

23.2% agent · top contributor 13%143 contributors

org-wide

Adoption is broad. Many contributors are committing agent-authored work, which is what a real organisational rollout looks like.

3,096

agent commits

10,264

human commits

143

agent contributors

13%

largest contributor

Assisted or autonomous

2497assisted

599autonomous

Assisted — a developer’s identity authored the commit and the agent added itself as co-author. The agent ran git commit during a working session, with a person there. Autonomous — a bot identity is the author, no human in the loop. Across the whole index, 93.7% of agent commits are assisted.

Agent share by month

Jan26

Feb26

Mar26

Apr26

May26

Jun26

Jul26

Aug26

Sep26

Peak month: 35.4%. Bars are scaled to that peak, not to 100%.

Which agents

Devin	Cognition	1377	10.3%
Claude Code	Anthropic	1260	9.4%
Cursor Agent	Anysphere	499	3.7%
GitHub Copilot	GitHub	26	0.2%
OpenAI Codex	OpenAI	22	0.2%
Other / generic agent	Various	5	0%

Verify it yourself

Sample commits that matched. Every one links to GitHub — check the trailers.

65f9d9bbtest(proxy): use the module-level HTTPException import in the /nvidia_nim Devin
c9dd4b44revert(model_info): keep Gemini baseline majors single-digitDevin
945adfb6refactor(model_info): support multi-digit Gemini majorsDevin
501be314fix(proxy): enforce organization budgets when max_budget is 0Claude Code
eddfb5fbfix(responses): preserve hosted web search calls (#40828)Claude Code
55c5babbdocs(pr-template): fold assumptions guidance into Caveats instructionsClaude Code
f4e21430test(auth): freeze prefetch cache clock so the org getter cannot miss on aCursor Agent
75b16df6fix(auth): keep prefetched org entries on the 5s getter TTLCursor Agent
abaa2f8bfix(auth): put TQ008 suppressions on the patch call linesCursor Agent
5677bc23fix(proxy/batches): resolve managed unified input_file_id to storage_url wGitHub Copilot
0a17c7c3feat: add LiteLLM Rust workspace with Mistral OCR bridge (#31033)GitHub Copilot
1ccc1e5bchore: litellm oss staging160626 (#30527)GitHub Copilot
cc400502fix(mcp): canonicalize bearer scheme on bridge egressOpenAI Codex
1a2a24ecchore(mcp): document mutable bridge header shapeOpenAI Codex
da036ad0fix(mcp): harden DCR bridge admissionOpenAI Codex
b84f7f82Litellm oss staging (#29492)Other / generic agent
e8fcb012Litellm OSS Staging (#29161)Other / generic agent
58cd6f68test: fix misleading precedence test per review feedbackOther / generic agent

What was excluded

Commits in window	18,090
Merge commits removed	−4,704
CI / dependency bots removed	−26
Considered	13,360

Bots are excluded from the denominator entirely. Counting dependabot as “AI-written code” is the fastest way to produce a meaningless number.

How this was measured

Measured over February–August 2026, cloned directly. Detection is deterministic string matching against signatures agents write about themselves — no heuristics, no style analysis. Merge commits and CI bots are excluded from the denominator. The figure is a floor: squash merges strip trailers, so real usage is higher. Contributor counts are aggregate — no individual is named.

Run it yourself: npx qmmit-cli BerriAI/litellm

Scan another repository

Try