Llama 3.1 70B Instruct RM RB2: benchmark results
Provider: Meta. Access: Open.
Unified ELO 1768 ± 44, rank #125 of 1605 rated models, from 17 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| RewardBench Ties | 88.35 | Score (%) | 97.9 |
| RewardBench Factuality | 81.26 | Accuracy (%) | 96.9 |
| RewardBench | 90.21 | Score (%) | 92.6 |
| RewardBench Math | 69.95 | Score (%) | 88.3 |
| RewardBench Safety | 90.95 | Accuracy (%) | 88.3 |
| RewardBench Chat Hard | 83.55 | Accuracy (%) | 86.4 |
| RewardBench Precise IF | 41.88 | Score (%) | 83.2 |
| Themis-CodeRewardBench - Functional Correctness | 86.23 | Preference accuracy (%; share of the benchmark's preference | 80.6 |
| RewardBench Focus | 86.46 | Score (%) | 79.6 |
| RewardBench Chat | 96.65 | Accuracy (%) | 77.8 |
| Themis-CodeRewardBench | 78.23 | Preference accuracy (%; share of the benchmark's preference | 76 |
| Themis-CodeRewardBench - Readability and Maintainability | 71.25 | Preference accuracy (%; share of the benchmark's preference | 74 |
Interactive version: theaggregate.ai/model?slug=llama-3-1-70b-instruct-rm-rb2 · How It Works · Data refreshed daily, snapshot 2026-09-26.