Score breakdown
Popularity is tracked separately. Support, ads, sponsorships, and tips never affect these signals.
Why it matters
Autoresearch belongs on RepoRadar because it ships a real, runnable autonomous-research primitive rather than a chat wrapper or skill collection. The agent edits the actual training code (`train.py`) -- architecture, optimizer, hyperparameters, batch size, everything -- not a config file or a `program.md` wrapper. The 5-minute wall-clock budget per experiment makes the loop tractable on a single
Who should use it
Who should skip it
Skip karpathy/autoresearch unless the captured evidence suggests it solves a problem you are actively working on.
About this signal
karpathy/autoresearch is tracked by RepoRadar as a code repository in the AI tooling section. It was first seen on 2026-07-12 and last updated on 2026-07-12. The current verdict is 'try now' with a Gold tier and moderate setup difficulty. Across RepoRadar's eight signals, karpathy/autoresearch is strongest on workflow potential (9.6) and momentum (9.5) and weakest on setup ease (6.4) — a profile worth weighing against your own priorities. This page summarizes the public evidence on the linked source page and states where additional review is still needed. The score, tier, risk label, and verdict on this page are never influenced by sponsorship, ads, or tips — they reflect only the usefulness, popularity, novelty, momentum, maturity, and evidence signals described in the RepoRadar methodology.
How this item is evaluated
RepoRadar assigned karpathy/autoresearch a composite score of 8.5 out of 10, placing it in the Gold tier. This score combines weighted sub-signals: usefulness (35%), novelty (18%), momentum (14%), maturity (10%), open-source/build quality (7%), evidence quality (6%), workflow potential (6%), and setup ease (4%). Popularity is tracked separately at 100.0 and never affects the composite score or tier. The risk label of 'low' reflects inherent user-impacting hazards, not generic novelty. Items with no risk flag may still require normal code review before production use.
Putting this into practice? Read How to read AI benchmarks without getting fooled for the checklist behind this score.
Risk explanation
Last commit was 2026-03-26 (~3.5 months ago) -- under the cycle 141 4-month stale threshold but approaching it; the cycle-script chose the pick on the strength of the README's authoritative MIT declaration and Karpathy's MIT-by-convention track record (nanochat, llm.c, minbpe are all MIT); No LICENSE file at the repo root (raw main/LICENSE returns 404); the MIT declaration lives only in the README's `## License / MIT` section -- authoritative for the maintainer's intent but weaker than a dedicated LICENSE file for downstream legal review; Single-machine scope: the README targets a single H100 (or comparable) GPU; multi-GPU / cluster workflows require the maintainer's separate `nanochat` project (56,175★) or the active community forks; Default-branch master is unusual for a Python project in 2026 (most Python projects default to main); not a risk per se but worth noting for tooling that branches off.