Open Models Almanac
Methodsevaluation

single mention · evaluation

Standardized agent evaluation protocol

Applies shared sample counts, sampling parameters, context limits, and maximum steps across scaffolds.

The sources on this page only mention it in passing. The line above is the extraction's one-line gloss. See evaluation for the methods filed beside it.

Filed under evaluation