reasoning benchmark
CorpusQA
CorpusQA is a multi-document, free-form long-context question answering benchmark in which a model must retrieve and reason over information distributed across a large corpus to produce open-ended answers that are scored by an LLM judge.
Updated Aug 7, 2026
Published models1
Registry coverage1
MetricScore
EvidenceB
CorpusQA leaderboard
Sorted by the source-provided rank. Higher score is better according to the registry.
| 1 | MI | 82.0% | 100.0% | 1 | C |
CorpusQA highlights
The top published results on this benchmark's own scale.
What is CorpusQA?
Definition and scoring fields from the benchmark registry.
CorpusQA is a multi-document, free-form long-context question answering benchmark in which a model must retrieve and reason over information distributed across a large corpus to produce open-ended answers that are scored by an LLM judge.
Scores are shown in ratio. The current registry marks this benchmark as not independently verified with evidence level B.
- Family
- CorpusQA
- Modality
- text
- Primary category
- reasoning
- Score direction
- higher
- LLMBoard eligible
- No
- Evaluation key
- corpusqa|llm-stats-current
Source-native results are preserved. Eligibility for the overall LLMBoard score is a separate policy decision.
FAQ
Common questions about CorpusQA.
Which model scores highest on CorpusQA?
MAI-Thinking-1 is currently ranked first with 82.0%.
What does CorpusQA measure?
CorpusQA is a multi-document, free-form long-context question answering benchmark in which a model must retrieve and reason over information distributed across a large corpus to produce open-ended answers that are scored by an LLM judge.
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
1 unique published model results are currently shown.
Does this benchmark affect the overall score?
This benchmark is preserved as source-native evidence but is not eligible for the current overall score.