Sakura-SOLRCA-Math-Instruct-DPO-v2: benchmark results
Sakura version-2 DPO on combined Intel Orca and Argilla math preferences. Similar recipe does not establish identity with version 1. Provider: Other. Access: Open.
Aggregate Rating 1461 ± 38, rank #3293 of 10916 rated models, from 6 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open LLM Leaderboard v1 - TruthfulQA MC2 | 72.16 | MC2 (%) (0-shot) | 93.2 |
| Open LLM Leaderboard v1 - HellaSwag | 88.52 | Normalized accuracy (%) (10-shot) | 92.6 |
| Open LLM Leaderboard v1 - ARC Challenge | 71.25 | Normalized accuracy (%) (25-shot) | 88.6 |
| Open LLM Leaderboard v1 - WinoGrande | 83.03 | Accuracy (%) (5-shot) | 86.8 |
| Open LLM Leaderboard v1 - MMLU | 66.13 | Accuracy (%) (5-shot) | 86 |
| Open LLM Leaderboard v1 - GSM8K | 63.91 | Accuracy (%) (5-shot) | 77.6 |
Interactive version: theaggregate.ai/model?slug=sakura-solrca-math-instruct-dpo-v2 · How It Works · Data refreshed daily, snapshot 2026-11-08.