Skip to main content
ModelScale

external benchmark

Terminal-Bench Science (Vals) leaderboard

Vals Terminal-Bench Science 0.1. Every model the catalog carries a published Terminal-Bench Science (Vals) value for, ranked by that value.

Categoryexternal
MeasureStrict pass@1 accuracy
Tasks70 scientific research workflows across five domains
DifficultyExpert scientific research workflows

Vals AI’s fixed-harness, single-trial run of 70 expert-authored scientific research workflows.

No model in the catalog has a published Terminal-Bench Science (Vals) score.

The benchmark is defined by Vals Terminal-Bench Science, but the catalog carries no value for it yet. An absent value is shown as absent here rather than as a zero.

All leaderboards