RRM-32B: benchmark results
Provider: Other.
Aggregate Rating 1588 ± 39, rank #1146 of 10916 rated models, from 12 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Plan-RewardBench - Multi-Turn Planning (Hard) | 62.1 | Pairwise accuracy (%) on the 73 hard multi-turn planning pai | 60 |
| Plan-RewardBench - Multi-Turn Planning (Easy) | 68.45 | Pairwise accuracy (%) on the 109 easy multi-turn planning pa | 55 |
| Plan-RewardBench - Safety Refusal | 60.3 | Pairwise accuracy (%) on the 51 safety-refusal pairs of the | 55 |
| IF-RewardBench - Overall Assessment (Single-Turn) | 0.14 | Kendall tau-b (-1 to 1) between the judge ranking of the res | 50 |
| Plan-RewardBench - Error Recovery | 67.15 | Pairwise accuracy (%) on the 361 robust error-recovery pairs | 50 |
| Plan-RewardBench - Single-Turn Planning (Hard) | 70.8 | Pairwise accuracy (%) on the 158 hard single-turn planning p | 50 |
| Plan-RewardBench - Tool Irrelevance | 61.15 | Pairwise accuracy (%) on the 275 tool-irrelevance or tool-un | 50 |
| Plan-RewardBench - Single-Turn Planning (Easy) | 75.22 | Pairwise accuracy (%) on the 144 easy single-turn planning p | 45 |
| IF-RewardBench - Overall Assessment | 0.07 | Kendall tau-b (-1 to 1) between the judge ranking of the res | 37.5 |
| Plan-RewardBench | 66.45 | Pairwise accuracy (%), macro-average of the seven scenario c | 35 |
| IF-RewardBench - Overall Assessment (System-Prompt) | 0 | Kendall tau-b (-1 to 1) between the judge ranking of the res | 25 |
| IF-RewardBench - Overall Assessment (Multi-Turn) | 0.07 | Kendall tau-b (-1 to 1) between the judge ranking of the res | 12.5 |
Interactive version: theaggregate.ai/model?slug=rrm-32b · How It Works · Data refreshed daily, snapshot 2026-11-08.