Hateful content detection on Hateful Memes
82.3AccuracyOpus-4.6
Evaluation Results
| Method | Links | |
|---|---|---|
| Opus-4.6Model Scale=≈800B2026.06 | 82.3 | |
| K2.5Model Scale=≈1T2026.06 | 81.4 | |
| Yuvion VLModel Scale=32B, Decoding Mode=Instruct2026.06 | 79.8 | |
| Qwen3.5-PlusModel Scale=397B2026.06 | 78.6 | |
| Yuvion VLModel Scale=32B, Decoding Mode=Reasoning2026.06 | 78 | |
| Qwen3-VLModel Scale=32B, Decoding Mode=Thinking2026.06 | 76.3 | |
| Yuvion VLModel Scale=8B, Decoding Mode=Instruct2026.06 | 76.2 | |
| GPT-5.4Model Scale=≈1.5T2026.06 | 74 | |
| Qwen3-VLModel Scale=8B, Decoding Mode=Thinking2026.06 | 73.3 | |
| Yuvion VLModel Scale=8B, Decoding Mode=Reasoning2026.06 | 72.9 | |
| Qwen3-VLModel Scale=32B, Decoding Mode=Instruct2026.06 | 71.5 | |
| Qwen3-VLModel Scale=8B, Decoding Mode=Instruct2026.06 | 68.8 |