Methodology
How Meridias evaluates each capability category
Meridias independently evaluates vendor APIs and open-source models within each capability category using proprietary datasets, category-specific metrics and consistent test conditions. Evaluations are repeated continuously as vendor APIs and open-source models change.
Independent evaluation
Commercial relationships do not influence evaluation results.
Methodology publications
We will publish a methodology paper for each capability category, explaining its scope, datasets, metrics, test conditions and evaluation process. Proprietary datasets, scores and rankings will remain private. More information is coming soon.