ResearchBenchmarksVisual Question Answering on VQA Reliable Challenge v2 (test-standard)For AI assistantsMarkdownContent indexVisual Question Answering on VQA Reliable Challenge v2 (test-standard)Follow84.11AccuracyDAVR81.239681.984882.7383.4752Dec 16, 2025Evaluation resultsMethodMethodLinksAccuracyPhi100AUC (100)Coverage @ 0.5%Coverage @ 1%Coverage @ 2%Coverage @ 5%DAVR2025.1284.1139.6497.2245.6754.1263.7178.55jokur2025.1284.0532.0696.8540.6749.359.1575.3TIMM2025.1283.9131.3196.6337.9346.4656.3872.86Freedom2025.1282.9922.7396.7229.2144.7158.8476.65Baseline2025.1281.352396.0926.9839.7552.8871.78