Instruction Following Evaluation on InfoBench
86.1ScoreRM-R1-32B (Qwen-2.5-Inst)
Evaluation results
| Method | Links | |
|---|---|---|
| RM-R1-32B (Qwen-2.5-Inst)Backbone=Qwen-2.5-Inst2026.05 | 86.1 | |
| Gemini-2.5-FlashModel Category=Black-box LLMs2026.05 | 85.6 | |
| RM-R1-14B (Qwen-2.5-Inst)Backbone=Qwen-2.5-Inst2026.05 | 85.5 | |
| RM-R1-32B (DeepSeek-Dist)Backbone=DeepSeek-Dist2026.05 | 85 | |
| RRM-32BModel Category=Larger White-box LLMs2026.05 | 84.4 | |
| Rubric-ARROW-voting@5Voting Strategy=voting@52026.05 | 84 | |
| Rubric-RM-voting@5Voting Strategy=voting@52026.05 | 83.8 | |
| RM-R1-14B (DeepSeek-Dist)Backbone=DeepSeek-Dist2026.05 | 82.4 | |
| API Prompting (Rubric+Judge pairwise)Evaluation Protocol=Rubric+Judge pairwise2026.05 | 82.2 | |
| Rubric-ARROWModel Category=Rubric-based Methods2026.05 | 82.1 | |
| Rubric-RMModel Category=Rubric-based Methods2026.05 | 80.8 | |
| RIFL-voting@5Voting Strategy=voting@52026.05 | 80.1 | |
| Rubric-ARROW w/o RL-voting@5Voting Strategy=voting@5, Training Protocol=w/o RL2026.05 | 79.7 | |
| RIFLModel Category=Rubric-based Methods2026.05 | 78.4 | |
| Rubric-ARROW w/o RLTraining Protocol=w/o RL2026.05 | 75.9 | |
| Qwen-3-8B (Rubric+Judge pairwise)Backbone=Qwen-3-8B, Evaluation Protocol=Rubric+Judge pairwise2026.05 | 74.6 | |
| API Prompting (direct Judge)Evaluation Protocol=direct Judge2026.05 | 72.9 | |
| RM-R1-7B (Qwen-2.5-Inst)Backbone=Qwen-2.5-Inst2026.05 | 71.3 | |
| RM-R1-7B (DeepSeek-Dist)Backbone=DeepSeek-Dist2026.05 | 70.3 | |
| Qwen-3-8B (Rubric+Judge pointwise)Backbone=Qwen-3-8B, Evaluation Protocol=Rubric+Judge pointwise2026.05 | 69.9 | |
| RRM-7BModel Category=White-box Judge/Reward LLMs2026.05 | 68.2 | |
| JudgeLRM-7BModel Category=White-box Judge/Reward LLMs2026.05 | 62.7 | |
| API Prompting (Rubric+Judge pointwise)Evaluation Protocol=Rubric+Judge pointwise2026.05 | 44.8 |