Inference
EN

Discrete Reasoning Over Paragraphs

A reading-comprehension benchmark requiring discrete reasoning over paragraphs, reported in DeepSeek-V4 base-model evaluations.

  • Not in index
  • Reasoning
ModelsModels with a result.6
Top resultBest result on this benchmark.—
Top-3 spreadPoints from first to third.—
YearYear of release.—

Results

6 results
#ModelResultThe score from the source.
—o1-mini83.9%
—GPT-4.583.4%
—Claude 3.5 Haiku83.1%
—Celeris-181.4%
—o4-mini77.7%
—Soofi S 30B-A3B66.5%