Multimodal Reward Modeling on RewardBench 2
96.7Safety ScoreSW-RM-V2-LLaMA3.1-8B
Evaluation Results
| Method | Links | |
|---|---|---|
| SW-RM-V2-LLaMA3.1-8B2026.01 | 96.7 | |
| UniRM2026.01 | 91.85 | |
| SW-VL-RM-7B2026.01 | 89.1 | |
| SW-RM-V2-Qwen0.6B2026.01 | 84.4 | |
| GuardRank2026.01 | 65.48 |
| Method | Links | |
|---|---|---|
| SW-RM-V2-LLaMA3.1-8B2026.01 | 96.7 | |
| UniRM2026.01 | 91.85 | |
| SW-VL-RM-7B2026.01 | 89.1 | |
| SW-RM-V2-Qwen0.6B2026.01 | 84.4 | |
| GuardRank2026.01 | 65.48 |