Skip to content
CAI
Produce a survey ↗Verify a survey

karpathy/nanoGPT

This system is a toolkit for training and evaluating GPT-2 style language models, specifically focusing on character-level and base model fine-tuning. It provides scripts to prepare datasets like Shakespeare and OpenWebText, along with configuration files for training and evaluation. The repository also includes utilities for benchmarking model performance and analyzing scaling laws.

39.6

Weak · 2 August 2026

1.1k

lines of production code

Python

primary language

1

bus factor · 38 authors in all

3

measurements over time

CAI band scale
CAI trend line
CAI lens gauges

Survey your own repository

karpathy/nanoGPT was measured the same way every project in this corpus was: the same rubric, at a pinned commit, with the result published in full. Point the surveyor at a repository you know and see whether you agree with it.

Survey a repository

About this page

  • The description of this project is derived from its own commit history, not from its README.
  • The score is its highest published measurement, taken on 2 August 2026 at a pinned commit. It is not a live figure and does not change until the project is measured again.
  • Measured at commit 3adf61e154 — the exact code this score is about.
  • Scored under rubric rubric-2026.08.18. Score the same commit under that rubric and you get the same number.
CAI link cards