How teams that share their AEI results score
The median and the middle half of the shared results, overall and for each dimension.
The site holds 0 shared results today.
No comparison is shown yet: a figure is shown only for a group of at least 11 results.
The groups
| Group | Shared results | Overall score |
|---|---|---|
| All shared results | 0 | no comparison: 0 results, fewer than 11 |
How to read the figures
A figure is shown only for a group of at least 11 results. The middle half runs from the lower quartile to the upper quartile: half of the group's scores are inside it. The shared results are the results whose owners chose to share them, 1 per member and coding agent. They are not a sample of all teams.
The site computes the figures again at each visit, from the results shared at that moment that were graded with the check catalog of 2026-11-16. A shared result graded with an earlier catalog covers other checks and is left out until its owner uploads a new run. The site publishes no mean, no lowest and no highest score, no file name, no reason and no owner.
To add your result, sign in, open the result and turn sharing on. How the benchmark is built, and what AEI Scan finds in public repositories. The same figures for a program: benchmark.json.
Prosights.ai