M3 Hate Memes (Meme + Post): leaderboard
Metric: Binary hate-detection accuracy (%) on all 2,455 M3 memes (1,318 hate, 1,137 normal) collected from X, 4chan and Weibo; the model sees the meme image and the accompanying social-media post text; zero-shot; higher is better. Source: arxiv.org. Saturation forecast: Not forecast. 8 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 2.5 VL 7B Instruct | 86.48 | #643 |
| 2 | Qwen 3 VL 8B Instruct | 85.95 | #401 |
| 3 | GPT-4o | 85.47 | #333 |
| 4 | GLM-4.1V-9B (Thinking) | 77.11 | #457 (GLM-4.1V-9B) |
| 5 | Gemini 3 Pro | 73.44 | #77 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #8 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=m3-hate-memes-meme-plus-post · How It Works · Data refreshed daily, snapshot 2026-10-11.