aider-polyglot@1.0
Coding
A polyglot coding benchmark that evaluates AI agents’ ability to perform code editing and generation tasks across multiple programming languages.
Run this task
CLI:
inspect eval inspect_harbor/aider_polyglot_1_0 --model openai/gpt-5Python:
from inspect_ai import eval
from inspect_harbor import aider_polyglot_1_0
eval(aider_polyglot_1_0(), model="openai/gpt-5")Dataset information
| Harbor registry | aider-polyglot@1.0 |
| Inspect task | aider_polyglot_1_0 |
| Version | 1.0 |
| Samples | 225 |
| Paper | arxiv |
See Task Parameters for the parameter set shared across all Harbor tasks.