OpenBookQA
text
+
+
+
+
About
OpenBookQA is a question-answering benchmark modeled after open-book exams, requiring models to combine elementary science facts with common sense reasoning. Unlike traditional QA tasks, it provides a knowledge book but requires multi-step reasoning to connect facts and arrive at answers, testing both retrieval abilities and logical inference skills.
+
+
+
+
Evaluation Stats
Total Models4
Organizations2
Verified Results0
Self-Reported4
+
+
+
+
Benchmark Details
Max Score1
Language
en
+
+
+
+
Performance Overview
Score distribution and top performers
Score Distribution
4 models
Top Score
89.6%
Average Score
77.2%
High Performers (80%+)
1Top Organizations
#1Microsoft
3 models
82.7%
#2Mistral AI
1 model
60.6%
+
+
+
+
Leaderboard
4 models ranked by performance on OpenBookQA
| License | Links | ||||
|---|---|---|---|---|---|
| Aug 23, 2024 | MIT | 89.6% | |||
| Aug 23, 2024 | MIT | 79.2% | |||
| Feb 1, 2025 | MIT | 79.2% | |||
| Jul 18, 2024 | Apache 2.0 | 60.6% |
