Pulse
2026-09-18-v3.0
Dataset
2026-10-27-catalog-ai-v2
Ruleset
b2e8889c66f7
Scanned
2026-10-27
Window
February–September 15, 2026
Catalog tracking observation. This repository is measured with the same fixed window and ruleset, but it is not part of the published benchmark or its aggregate claims.

Scalable toolkit for efficient model reinforcement

01Result

NVIDIA-NeMo/RL ↗

NVIDIA-NeMo/RL AI coding statistics: 23.5%

of 776 commits in February–September 15, 2026 were committed by an AI agent — a floor, not an estimate
23.5% agent · top contributor 20.9%49 contributors
org-wide

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

182
agent commits
594
other considered commits
49
agent contributors
20.9%
largest contributor

02Drivers

What shapes this result

Mode, timing, and detected signatures show whether the result reflects broad adoption or one concentrated workflow.

Assisted or autonomous

182assisted
0autonomous

Assisted — a developer’s identity authored the commit and the agent added itself as co-author. Autonomous — a bot identity authored the commit without a human identity attached.

Agent share by month

Agent-attributed commit share by month
MonthAgent shareConsidered commits
2026-024.7%64
2026-033.3%60
2026-0424.2%66
2026-0530.9%81
2026-0632%150
2026-0727%148
2026-0827.3%139
2026-0914.7%68

Peak month: 32%. Bars are scaled to that peak, not to 100%.

Which agents

AgentVendorCommitsShare
Claude CodeAnthropic17822.9%
Cursor AgentAnysphere40.5%
DevinCognition10.1%
GitHub CopilotGitHub10.1%

03Evidence

Inspect the matches

Public commit samples let you inspect the signatures behind the aggregate result.

Matched commit samples

  • c69a6427 on GitHubfix(megatron): synchronize parameters before inference refit (#4086)Claude Code
  • 53bce056 on GitHubfeat(sc): restart dead generation shards and re-admit them at the next refClaude Code
  • c49d53e2 on GitHubfeat(mopd): add full-vocabulary on-policy distillation to SingleControllerClaude Code
  • 4454fa64 on GitHubfeat: data plane transfer queue integration (#2439)Cursor Agent
  • e7266a92 on GitHubfix: fix skip_reference_policy_logprobs_calculation and skip_prev_logprobsCursor Agent
  • 2841fefb on GitHubfeat: ProRLv2 - add seq-mask-tis truncated importance sampling type (#1899Cursor Agent
  • 81115337 on GitHubfeat: add DeepSeek V4 Flash GRPO support - Automodel path (#3595)Devin
  • f0185566 on GitHubfeat: Expose dp_replicate_size for hybrid FSDP in Automodel DTensor v2 (#2GitHub Copilot

04Method

What entered the denominator

The result begins with commit history, removes merges and routine automation, and reports literal agent signatures as a lower bound.

Exclusion ledger

Commits in window776
Merge commits removed−0
CI / dependency bots removed−0
Considered776

Bots are excluded from the denominator entirely. Agent attribution requires a literal supported signature; no code-style inference is used.

Continue exploring

Claude Code commitsBrowse repositories with detected Claude Code signatures →GitHub Copilot commitsBrowse repositories with detected GitHub Copilot signatures →Cursor Agent commitsBrowse repositories with detected Cursor Agent signatures →Devin commitsBrowse repositories with detected Devin signatures →Python AI coding statisticsCompare evidence across Python repositories →GitHub AI commit analyzerInspect another repository with the same evidence model →
Reproduce this observation

Run the same detector locally.

Deterministic signature matching; no code-style inference. Squash merges can remove trailers, so this result remains a measured floor.

npx qmmit-cli NVIDIA-NeMo/RL

Scan another repository

Try