← the index
vllm-project/vllm ↗
12.5%
of 7,908 commits were committed by an AI agent — a floor, not an estimate
12.5% agent · top contributor 3.6%464 contributors
org-wide

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

991
agent commits
6,917
human commits
464
agent contributors
3.6%
largest contributor

Assisted or autonomous

991assisted
0autonomous

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: 22.4%. Bars are scaled to that peak, not to 100%.

Which agents

Claude CodeAnthropic6758.5%
OpenAI CodexOpenAI1111.4%
Gemini CLIGoogle1091.4%
Cursor AgentAnysphere931.2%
GitHub CopilotGitHub320.4%
DevinCognition20%

Verify it yourself

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

  • 0fefffc9[KVConnector][NIXL] Support attention-HMA layouts in pipeline-parallel pusClaude Code
  • 12575123[DSA] Wire DeepGEMM sparse MQA logits into the DeepSeek V4.1 indexer (#562Claude Code
  • 1e47ec00[Bugfix][Gemma 4] Don't read fft_length when profiling unified audio (#567Claude Code
  • 031f5810[Build][NVIDIA] Update public Rubin dependencies and MSA compatibility (#5OpenAI Codex
  • 0fefffc9[KVConnector][NIXL] Support attention-HMA layouts in pipeline-parallel pusOpenAI Codex
  • 241e9391[Bugfix][Attention] Stabilize sparse-MLA DCP for GLM PCP evals (#55879)OpenAI Codex
  • 6f74337c[Bugfix][Spec Decode]Preserve user --speculative-config overrides for specGemini CLI
  • 8fea1d30Fix uniform_random routing simulation to sample without replacement (#4368Gemini CLI
  • bf2b45b5[Attention] Integrate FlashAttention 4 SM100 headdim 256 support (#42669)Gemini CLI
  • 35dc2730[ROCm][Spec Decode] Add Aiter MLA decode support non-causal draft block (#Cursor Agent
  • 79e205e8[Doc] Clarify ITL vs TPOT Prometheus metrics (#55283)Cursor Agent
  • 00972dfd[ROCm][DSV4.1][Perf] Fold the mHC post step into the delayed pre projectioCursor Agent
  • c9355e25[XPU][Bugfix] Add forward_xpu to Ernie4_5_VLRotaryEmbedding (#55942)GitHub Copilot
  • 6fd08c45[XPU][Bugfix] Fix FalconH1 pipeline-parallel execution (#55913)GitHub Copilot
  • a0d3e5c1[XPU] Add fused GemmaRMSNorm path for eager execution (#53678)GitHub Copilot
  • d3d79ffc[Bugfix][Spec Decode] Capture the widest uniform decode batch by default (Devin
  • 044b0522[Kernel][Perf] Tune fused_moe FP8 config for Qwen3.5 on L40S (+7%) (#53819Devin

What was excluded

Commits in window7,911
Merge commits removed−0
CI / dependency bots removed−3
Considered7,908

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 vllm-project/vllm

Scan another repository

Try