reasoning benchmark
Natural Questions
Natural Questions is a question answering dataset featuring real anonymized queries issued to Google search engine. It contains 307,373 training examples where annotators provide long answers (passages) and short answers (entities) from Wikipedia pages, or mark them as unanswerable.
Updated Aug 11, 2026
Models7
Model coverage7
MetricScore
EvidenceB
Natural Questions Ranking
Higher score ranks better on this benchmark.
| 01 | GO | 34.5% | 100.0% | 7 | C | |
| 02 | MA | 31.2% | 83.3% | 7 | C | |
| 03 | GO | 29.2% | 66.7% | 7 | C | |
| 04 | GO | 20.9% | 50.0% | 7 | C | |
| 05 | GO | 20.9% | 33.3% | 7 | C | |
| 06 | GO | 15.5% | 16.7% | 7 | C | |
| 07 | GO | 15.5% | 0.0% | 7 | C |
Natural Questions Score Distribution
A closer view of the leading scores on this benchmark.
Natural Questions Highlights
The leading models and scores on this benchmark.
What is Natural Questions?
What Natural Questions measures and how its scores work.
Natural Questions is a question answering dataset featuring real anonymized queries issued to Google search engine. It contains 307,373 training examples where annotators provide long answers (passages) and short answers (entities) from Wikipedia pages, or mark them as unanswerable.
Scores are shown in ratio. This benchmark is not independently verified and has an evidence level of B.
- Family
- Natural Questions
- Modality
- text
- Primary category
- reasoning
- Score direction
- higher
- LLMBoard eligible
- Yes
- Evaluation key
- natural-questions|llm-stats-current
Benchmark scores retain their original unit. Overall score eligibility is shown separately.
FAQ
Common questions about Natural Questions.
Which model scores highest on Natural Questions?
Gemma 2 27B is currently ranked first with 34.5%.
What does Natural Questions measure?
Natural Questions is a question answering dataset featuring real anonymized queries issued to Google search engine. It contains 307,373 training examples where annotators provide long answers (passages) and short answers (entities) from Wikipedia pages, or mark them as unanswerable.
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
7 model results are currently shown.
Does this benchmark affect the overall score?
Yes. This benchmark can contribute to the current LLMBoard capability score.