Aider
Aider is one of the original edit-by-diff coding agents and the harness behind the widely cited Aider Polyglot leaderboard. It runs locally, supports any model with an OpenAI-compatible API, and is famous for shipping working diff edits rather than full file rewrites, which makes it cheap to run on long sessions.
Leaderboard Placements
Live board as of 2026-10-28: federated Terminal-Bench rows, with TensorFeed editorial scores for harnesses the board does not cover.
| Benchmark | Best base model | Score | Rank |
|---|---|---|---|
| SWE-bench Verified | - | - | - |
| Terminal-Bench | Claude Opus 4.7 | 31.2 | #2 / 7 |
| Terminal-Bench 2.1 | - | - | - |
| Terminal-Bench 3.0 | - | - | - |
| Terminal-Bench 4.0 | - | - | - |
| Aider Polyglot | Claude Opus 4.7 | 84.2 | #1 / 3 |
| SWE-Lancer | - | - | - |
Distribution
Open-source Python CLI. Install via pip. Apache 2.0 license.
Model Story
Multi-model. Anthropic, OpenAI, Google, DeepSeek, and any OpenAI-compatible endpoint. Bring your own key.
Pricing
Free harness; you pay for the underlying API tokens.
Who It's For
Engineers who want a local-first, model-agnostic terminal agent and care about token efficiency on long edits.
Notable Features
- Edit-by-diff over whole-file rewrites
- Self-published Polyglot leaderboard (225 hardest Exercism tasks)
- Git-aware: every edit is a commit
- Voice mode
- Architect / Editor split-model routing