laser-dolphin-mixtral-2x7B-dpo: benchmark results
A mixture-of-experts release by macadeliccc combining cognitivecomputations/dolphin-2.6-mistral-7b-dpo-laser and teknium/OpenHermes-2.5-Mistral-7B. Its card credits Fernando Fernandes and Eric Hartford for laserRMT and specifies Dolphin-style ChatML prompting. Provider: macadeliccc. Access: Open.
Aggregate Rating 1421 ± 34, rank #5108 of 10916 rated models, from 6 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open LLM Leaderboard v1 - HellaSwag | 85.8 | Normalized accuracy (%) (10-shot) | 76.8 |
| Open LLM Leaderboard v1 - TruthfulQA MC2 | 60.76 | MC2 (%) (0-shot) | 76.5 |
| Open LLM Leaderboard v1 - ARC Challenge | 65.96 | Normalized accuracy (%) (25-shot) | 71.7 |
| Open LLM Leaderboard v1 - WinoGrande | 79.01 | Accuracy (%) (5-shot) | 66.7 |
| Open LLM Leaderboard v1 - GSM8K | 48.29 | Accuracy (%) (5-shot) | 61.5 |
| Open LLM Leaderboard v1 - MMLU | 63.17 | Accuracy (%) (5-shot) | 58.8 |
Interactive version: theaggregate.ai/model?slug=laser-dolphin-mixtral-2x7b-dpo · How It Works · Data refreshed daily, snapshot 2026-11-08.