MMBench_test

multimodal
+
+
+
+
About

MMBench-Test is the test split of the MMBench benchmark, specifically designed for final evaluation of vision-language models without data leakage concerns. It maintains the same rigorous evaluation standards as the main MMBench while serving as a held-out test set for reliable performance assessment of multimodal models across bilingual visual understanding tasks.

+
+
+
+
Evaluation Stats
Total Models1
Organizations1
Verified Results0
Self-Reported1
+
+
+
+
Benchmark Details
Max Score1
Language
en
+
+
+
+
Performance Overview
Score distribution and top performers

Score Distribution

1 models
Top Score
86.5%
Average Score
86.5%
High Performers (80%+)
1

Top Organizations

#1Alibaba Cloud / Qwen Team
1 model
86.5%
+
+
+
+
Leaderboard
1 models ranked by performance on MMBench_test
LicenseLinks
Aug 29, 2024
tongyi-qianwen
86.5%
+
+
+
+
Resources