Skip to main content
ModelScale

Reasoning benchmark

HellaSwag leaderboard

Every model the catalog carries a published HellaSwag value for, ranked by that value.

CategoryReasoning
MeasureExact match
TasksCommonsense completion questions
DifficultyCommonsense reasoning

A commonsense natural-language inference benchmark reported in DeepSeek-V4 base-model evaluations.

No model in the catalog has a published HellaSwag score.

The benchmark is defined by DeepSeek-V4 Technical Report, but the catalog carries no value for it yet. An absent value is shown as absent here rather than as a zero.

All leaderboards · Reasoning capability leaderboard