dragonwar-7B-alpha: benchmark results
maldv Dragonwar 7B Alpha is a fantasy-story continuation fine-tune trained through full-text windows, completion and ORPO, followed by a further rotated full-text epoch. Its publisher recommends story prefixes rather than a chat prompt. Provider: maldv. Access: Open.
Aggregate Rating 1350 ± 34, rank #7687 of 10916 rated models, from 6 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open LLM Leaderboard v1 - HellaSwag | 81.54 | Normalized accuracy (%) (10-shot) | 45.6 |
| Open LLM Leaderboard v1 - GSM8K | 20.32 | Accuracy (%) (5-shot) | 41.2 |
| Open LLM Leaderboard v1 - MMLU | 56.71 | Accuracy (%) (5-shot) | 39.8 |
| Open LLM Leaderboard v1 - ARC Challenge | 56.57 | Normalized accuracy (%) (25-shot) | 35.7 |
| Open LLM Leaderboard v1 - TruthfulQA MC2 | 43.73 | MC2 (%) (0-shot) | 24.6 |
| Open LLM Leaderboard v1 - WinoGrande | 70.09 | Accuracy (%) (5-shot) | 24.4 |
Interactive version: theaggregate.ai/model?slug=dragonwar-7b-alpha · How It Works · Data refreshed daily, snapshot 2026-11-08.