MMBench
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
MMBench English (dev)
86.34Overall Score
9
MMBench-Video (test)
1.63GPT-4 Score
9
MMBench v1.1
82.2Accuracy
9
MMBench (test)
64.3Score
9
MMBench
58Accuracy
8
MMBench CN
79.6GPT Score
8
MMBench EN
85.4GPT Score
8
MMBench General Knowledge Retention
81.18COM Score
8
MMBench
-29.3BD-Rate
8
MMBench Overall & Relation Reasoning
84.7Overall Accuracy
8
MMBench EN 1.1 (dev)
84.31Accuracy
8
MMBench
67.01Score
8
MMBench Chinese (test)
63.6Accuracy
8
MMBench 10 task sets 2 (unseen)
38.04PSNR (dB)
7
MMBench2 10 task sets (seen)
39.66PSNR (dB)
7
MMBench
88.6Score
7
MMB (MMBench) (test)
84.1MMB Score
7
MMBench EN V1.1 (test)
80.68Accuracy
7
MMBench EN
85Score
7
MMBench
71.56Score
7
MMBench cn
89.7Pass@1
7
MMBench en
90.1Pass@1
7
MMBench-CN (test)
88.5Overall Score
7
MMBench (dev)
64.6Accuracy
6
MMBench (test)
64.7Accuracy
6