MiquMaid-v1-70B: benchmark results
An experimental Miqu role-play fine-tune by NeverSleep, credited to Undi and IkariDev. The card uses Alpaca prompting; version 1 remains separate from version 2 and its DPO derivatives. Provider: NeverSleep. Access: Open.
Aggregate Rating 1501 ± 33, rank #2352 of 10916 rated models, from 6 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open LLM Leaderboard v1 - WinoGrande | 85.08 | Accuracy (%) (5-shot) | 97.1 |
| Open LLM Leaderboard v1 - MMLU | 74.9 | Accuracy (%) (5-shot) | 96.3 |
| Open LLM Leaderboard v1 - ARC Challenge | 71.67 | Normalized accuracy (%) (25-shot) | 89.9 |
| Open LLM Leaderboard v1 - HellaSwag | 87.96 | Normalized accuracy (%) (10-shot) | 88.9 |
| Open LLM Leaderboard v1 - GSM8K | 69.29 | Accuracy (%) (5-shot) | 88.6 |
| Open LLM Leaderboard v1 - TruthfulQA MC2 | 61.79 | MC2 (%) (0-shot) | 78.2 |
Interactive version: theaggregate.ai/model?slug=miqumaid-v1-70b · How It Works · Data refreshed daily, snapshot 2026-11-08.