math benchmark
DROP
DROP (Discrete Reasoning Over Paragraphs) is a reading comprehension benchmark requiring discrete reasoning over paragraph content. It contains crowdsourced, adversarially-created questions that require resolving references and performing discrete operations like addition, counting, or sorting, demanding comprehensive paragraph understanding beyond paraphrase-and-entity-typing shortcuts.
Updated Aug 7, 2026
Published models30
Registry coverage30
MetricScore
EvidenceB
DROP leaderboard
Sorted by the source-provided rank. Higher score is better according to the registry.
Score distribution
Top published rows on the benchmark's original scale.
DROP highlights
The top published results on this benchmark's own scale.
What is DROP?
Definition and scoring fields from the benchmark registry.
DROP (Discrete Reasoning Over Paragraphs) is a reading comprehension benchmark requiring discrete reasoning over paragraph content. It contains crowdsourced, adversarially-created questions that require resolving references and performing discrete operations like addition, counting, or sorting, demanding comprehensive paragraph understanding beyond paraphrase-and-entity-typing shortcuts.
Scores are shown in ratio. The current registry marks this benchmark as not independently verified with evidence level B.
- Family
- DROP
- Modality
- text
- Primary category
- math
- Score direction
- higher
- LLMBoard eligible
- Yes
- Evaluation key
- drop|llm-stats-current
Source-native results are preserved. Eligibility for the overall LLMBoard score is a separate policy decision.
FAQ
Common questions about DROP.
Which model scores highest on DROP?
DeepSeek-V3 is currently ranked first with 91.6%.
What does DROP measure?
DROP (Discrete Reasoning Over Paragraphs) is a reading comprehension benchmark requiring discrete reasoning over paragraph content. It contains crowdsourced, adversarially-created questions that require resolving references and performing discrete operations like addition, counting, or sorting, demanding comprehensive paragraph understanding beyond paraphrase-and-entity-typing shortcuts.
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
30 unique published model results are currently shown.
Does this benchmark affect the overall score?
This benchmark is marked as eligible for the current LLMBoard capability methodology.