karpathy/nanoGPT

At a glance

Rank

#401 of 401

#56 of 56 Python repos

Checks passed

2 of 17

Full list below

Best set up for

Codex

Score 19.6

Needs most work for

Aider

Biggest fix: .aider.conf.yml (+14.1 pts)

What to fix first

Showing fixes for Claude Code

1. 1
   AGENTS.md / CLAUDE.md

   Add an AGENTS.md covering project goals, layout, setup commands, and conventions. Aim for 800+ chars of real guidance (not boilerplate).
   +13.2 pts
2. 2
   Test suite

   Add a tests/ (or test/, __tests__/, spec/) directory with runnable tests. Document how to run them in AGENTS.md.
   +13.2 pts
3. 3
   Reproducible dev env

   Add a Makefile or devcontainer or Dockerfile so the agent can set up the project in one command.
   +11.8 pts

Score for each AI agent

Each agent weighs the checks by its own docs. Under each score are its top three. How each agent is scored →

Does karpathy/nanoGPT have agent instructions?

An AGENTS.md (or CLAUDE.md) at the root tells AI coding agents how to build, test and work in the repo. A few agents also read a file of their own.

AGENTS.md / CLAUDE.md

✗None at the root (no AGENTS.md, CLAUDE.md, AGENT.md, .claude/CLAUDE.md or .claude/AGENTS.md)

Agent-specific files

None: no GEMINI.md, .cursor/rules, .openhands/setup.sh, .aider.conf.yml or .agents/skills

Repos that have one

29 of 56 Python repos · 221 of 401 repos on the leaderboard · See examples →

How karpathy/nanoGPT compares

vs leaderboard average

48.6 pts below

Average 66.8

vs Python average

52.8 pts below

Average 71.0

Stars among Python repos

#20 of 56

One fix for the most agents: Test suite would raise the score for all 9 agents, by up to 14.1 points.

Similar repos

Same language and host as this repo, closest in overall score.

• numpy/numpy53.1
  Python · 33.3k ★
• sgl-project/sglang51.5
  Python · 38.4k ★
• Significant-Gravitas/AutoGPT51.3
  Python · 187.9k ★
• langchain-ai/langchain50.7
  Python · 149k ★
• karpathy/nanochat39.3
  Python · 60.2k ★
• microsoft/AI23.2
  Python · 2.9k ★