Aider, the open-source coding agent, writes most of its own code

The Aider project tracks how much of each release Aider itself wrote. Recent releases report figures as high as 88%. It is a public, line-level record of an agent building its own product.

Organization: Aider (open source)
Industry: Developer Tools
When: Aug 2025
Evidence: Company reported
Agents and tools: Aider

Reported outcome

Aider's release notes report that Aider wrote up to 88% of the code in a release (v0.86.0, August 2025).

up to 88% per release
Share written by Aider

Caveats

Stated by the company that uses the agent (blog, earnings call, interview). Not independently audited.

The project measures this itself, by counting lines in git blame. Lines are not the same as effort or difficulty.
The figure changes a lot from release to release.

What the agent does

Aider runs in the terminal and edits files in a git repository from chat instructions. The maintainer uses Aider to build Aider, then counts how many new lines in each release came from Aider commits.

Why this is useful

Most claims about "AI wrote X% of our code" have no public data behind them. Aider's numbers come from public git history, so anyone can check the method.

What to copy

If you want to know how much an agent helps, measure it from git, not from a survey.
Make the agent commit its own changes with a clear author tag, so tracking is automatic.

Starter prompt

Written by us. A good first task for any terminal coding agent.

/add src/{module}.py tests/test_{module}.py

Add input validation to {function_name}:
- Reject {bad_input_cases} with a ValueError and a clear message.
- Add one test per rejected case.
Run the tests. Commit with the message "Add input validation to {function_name}".