ERQA
multimodal
+
+
+
+
About
ERQA (Embodied Reasoning Question Answering) is a multimodal benchmark that evaluates AI models' spatial reasoning and physical understanding capabilities through questions about embodied interactions. This benchmark tests models' ability to reason about 3D environments, spatial relationships, and physical consequences of actions, bridging the gap between language understanding and real-world physical reasoning for robotics applications.
+
+
+
+
Evaluation Stats
Total Models3
Organizations1
Verified Results0
Self-Reported3
+
+
+
+
Benchmark Details
Max Score1
Language
en
+
+
+
+
Performance Overview
Score distribution and top performers
Score Distribution
3 models
Top Score
65.7%
Average Score
55.0%
High Performers (80%+)
0Top Organizations
#1OpenAI
3 models
55.0%
+
+
+
+
Leaderboard
3 models ranked by performance on ERQA
