Observed arrival · 2026-09-22
DecisionEval puts decision models on the record
An independent directory tracks 100 decision models, 164 open-source projects, 86 comparisons, and benchmark results with the working shown.
- For
- Developers evaluating local decision models
- Worth noticing
- The benchmark splits 2,000 decisions by teacher agreement, reporting 0.841 on converged labels and 0.591 on disputed ones.
Field notes
The site combines a model directory with repository-level comparisons, open alternatives, runtime guides, and workflow walkthroughs. Its benchmark discussion separates 1,188 decisions on which teacher models agreed from 812 disputed decisions, then reports different accuracy figures for the two groups. Model entries are described as carrying an evidence grade, while project listings identify what each project decides, which model it calls, and when it was last pushed.
Observed signals
Read the marks
Editorial observations of this landing page, not a rating.
One card from the complete issue