CFMS - Sarcasm Target Recognition: leaderboard

CFMS (Chinese Fine-grained Multimodal Sarcasm, 2026) collects Chinese social-media image-text posts annotated for sarcasm, its target and an explanation; models answer zero-shot. This board scores the named target of the sarcasm on the sarcastic test posts, judged equivalent to the annotation by GPT-4o. Accuracy in percent; higher is better.

Metric: Target accuracy (%) over the sarcastic test posts: the named target of the sarcasm judged semantically equivalent to the annotation by GPT-4o, zero-shot; higher is better. Source: arxiv.org. Saturation forecast: Estimated already saturated. 4 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1Gemini 2.5 Flash82.54#463
2GPT-4o71.03#751
3InternVL2.5-8B40.47#2370
4Qwen 2.5 VL 7B Instruct32.54#1902

No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=cfms-sarcasm-target-recognition · How It Works · Data refreshed daily, snapshot 2026-11-21.