OK-VQA

OK-VQA: Evaluates multimodal understanding across image, text, chart, diagram, or cross-modal reasoning tasks.

18rows
overall_accuracyprimary metric
2026-05-27sampled

Metadata

ID
ok_vqa
Category
Intelligence
Release
Unknown

Metrics

Overall Accuracy

Latest Results

Rows are parsed from the public OK-VQA leaderboard table. Primary score is Overall Accuracy.

Rank Subject Overall Accuracy Model Match Provenance Sampled
1 Prophet 61.11% — Imported 2026-05-27
2 PromptCap 60.4% — Imported 2026-05-27
3 REVIVE 58% — Imported 2026-05-27
4 KAT 54.41% — Imported 2026-05-27
5 PICa 48% — Imported 2026-05-27
6 CBM 47.9% — Imported 2026-05-27
7 MCAN 44.65% — Imported 2026-05-27
8 VLC-BERT 43.14% — Imported 2026-05-27
9 UnifER 42.13% — Imported 2026-05-27
10 MAVEx 41.37% — Imported 2026-05-27
11 KRISP 38.9% — Imported 2026-05-27
12 ConceptBERT 33.66% — Imported 2026-05-27
13 MUTAN + AN 27.84% — Imported 2026-05-27
14 MUTAN 26.41% — Imported 2026-05-27
15 BAN + AN 25.61% — Imported 2026-05-27
16 BAN 25.17% — Imported 2026-05-27
17 MLP 20.67% — Imported 2026-05-27
18 Q only 14.93% — Imported 2026-05-27