MMMUval
multimodal
+
+
+
+
About
MMMUVal is the validation dataset for the Massive Multi-discipline Multimodal Understanding benchmark, serving as the standard development set for multimodal model evaluation. It provides a controlled environment for model validation across diverse academic disciplines while maintaining the same rigorous standards as the full MMMU benchmark.
+
+
+
+
Evaluation Stats
Total Models2
Organizations2
Verified Results0
Self-Reported2
+
+
+
+
Benchmark Details
Max Score1
Language
en
+
+
+
+
Performance Overview
Score distribution and top performers
Score Distribution
2 models
Top Score
77.8%
Average Score
71.2%
High Performers (80%+)
0Top Organizations
#1Anthropic
1 model
77.8%
#2Alibaba Cloud / Qwen Team
1 model
64.5%
+
+
+
+
Leaderboard
2 models ranked by performance on MMMUval
License | Links | ||||
---|---|---|---|---|---|
Sep 29, 2025 | Proprietary | 77.8% | |||
Aug 29, 2024 | tongyi-qianwen | 64.5% |