Benchmark
OpenBookQA
OpenBookQA is a question-answering dataset modeled after open book exams for assessing human understanding. It contains 5,957 multiple-choice…
- Modality
- text
- Categories
- reasoning, general
- Openness
- unknown
- Source
- llm_stats
- Reported scores
- 5
Reported scores
llm_stats
| Model | Organization | Reported value | Reported | Evidence |
|---|---|---|---|---|
| Hermes 3 70B | Nous Research | 0.494 | 2024-08-15 | evidence |
| Mistral NeMo Instruct | Mistral | 0.606 | 2024-07-18 | evidence |
| Phi 4 Mini | Microsoft | 0.792 | 2025-02-01 | evidence |
| Phi-3.5-mini-instruct | Microsoft | 0.792 | 2024-08-23 | evidence |
| Phi-3.5-MoE-instruct | Microsoft | 0.896 | 2024-08-23 | evidence |
Scores are partitioned by the source that reported them and are never merged into a single cross-source ranking, because the sources measure different things and say so.