ResearchBenchmarksMultimodal Sarcasm Explanation on WITSFollow3.18Human Evaluation ScoreQwen2.5VL2.48322.66412.8453.0259Jan 28, 2026Evaluation ResultsMethodMethodLinksHuman Evaluation ScoreLLM Judge ScoreQwen2.5VL2026.013.182.43MuVaC2026.012.823.36EDGE2026.012.763.36KimiVL2026.012.571.97MAF2026.012.543.28MOSES2026.012.513.24