- Dataset
- 2026-10-27-catalog-ai-v2
- Ruleset
- b2e8889c66f7
- Scanned
- 2026-10-27
- Window
- February–September 15, 2026
:metal: TT-NN operator library, and TT-Metalium low level kernel programming model.
01Result
tenstorrent/tt-metal ↗tenstorrent/tt-metal AI coding statistics: 30.4%
Adoption is broad. Many contributors are committing agent-authored work, which is what a real organisational rollout looks like.
02Drivers
What shapes this result
Mode, timing, and detected signatures show whether the result reflects broad adoption or one concentrated workflow.
Assisted or autonomous
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
| Month | Agent share | Considered commits |
|---|---|---|
| 2026-02 | 14.5% | 884 |
| 2026-03 | 19.4% | 1100 |
| 2026-04 | 25.4% | 989 |
| 2026-05 | 27.2% | 872 |
| 2026-06 | 35.1% | 1183 |
| 2026-07 | 40.5% | 1132 |
| 2026-08 | 38.7% | 1107 |
| 2026-09 | 42.8% | 608 |
Peak month: 42.8%. Bars are scaled to that peak, not to 100%.
Which agents
| Agent | Vendor | Commits | Share |
|---|---|---|---|
| Claude Code | Anthropic | 1640 | 20.8% |
| GitHub Copilot | GitHub | 572 | 7.3% |
| Cursor Agent | Anysphere | 359 | 4.6% |
| OpenAI Codex | OpenAI | 4 | 0.1% |
03Evidence
Inspect the matches
Public commit samples let you inspect the signatures behind the aggregate result.
Matched commit samples
b8ef12d0on GitHub[skip ci] ci(blaze-gate): pass this commit's dev image to tt-blaze so kernClaude Code627f631eon GitHub[Feature] Align serialized tensor data to 64 bytes (#56434)Claude Codeceaf66a1on GitHub[LLK] [SAN] llk_pack_init num_tiles !== llk_matmul_pack ntiles :facepalm: Claude Code61e0a505on GitHub[silencer] Silence unused dfb_x_id warning in RMSNORM layernorm kernel (#5GitHub Copilotb580317don GitHub[tt-train] [CI] Split tt-train merge_gate tests into 3 groups (#56530)GitHub Copilotf27b04b8on GitHub[silencer] Migrate deprecated pydantic `@validator` to `@field_validator` GitHub Copilot518c09eeon GitHubXTTS-v2 P150 Bringup (#52877)Cursor Agenta473dd32on GitHub[Performance] BEVFormer: Move SCA rebatch to device (#55494)Cursor Agent294bd9a0on GitHub[repo-assist] fix(ttnn): clamp gather codegen readers to prevent L1 out-ofCursor Agentf8c78f9bon GitHub[Bug fix] Gate the async-ahead token keep on `supports_async_decode` (#545OpenAI Codex521dee76on GitHub[Feature] Add QB2 Llama 3.1-8B and weekly agentic research CI (#55922)OpenAI Codex5502be5fon GitHubAdd autonomous TTNN model bringup framework (#54570)OpenAI Codex
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 window | 7,886 |
| Merge commits removed | −4 |
| CI / dependency bots removed | −7 |
| Considered | 7,875 |
Bots are excluded from the denominator entirely. Agent attribution requires a literal supported signature; no code-style inference is used.
Continue exploring
Related coding-agent evidence
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 tenstorrent/tt-metal