Llama-3-Lumimaid-70B-v0.1-alt: benchmark results
A Llama 3 role-play release by NeverSleep, credited to Undi and IkariDev. The card describes a mixture of role-play and general instruction datasets and uses the Llama 3 prompting format. The 70B alternate release remains distinct from the regular 70B checkpoint. Provider: NeverSleep. Access: Open.
Aggregate Rating 1526 ± 35, rank #1996 of 10916 rated models, from 6 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open LLM Leaderboard v1 - GSM8K | 80.06 | Accuracy (%) (5-shot) | 99.7 |
| Open LLM Leaderboard v1 - MMLU | 80.03 | Accuracy (%) (5-shot) | 99.6 |
| Open LLM Leaderboard v1 - WinoGrande | 84.77 | Accuracy (%) (5-shot) | 95 |
| Open LLM Leaderboard v1 - ARC Challenge | 71.33 | Normalized accuracy (%) (25-shot) | 89 |
| Open LLM Leaderboard v1 - HellaSwag | 86.28 | Normalized accuracy (%) (10-shot) | 79.9 |
| Open LLM Leaderboard v1 - TruthfulQA MC2 | 58.81 | MC2 (%) (0-shot) | 72.7 |
Interactive version: theaggregate.ai/model?slug=llama-3-lumimaid-70b-v0-1-alt · How It Works · Data refreshed daily, snapshot 2026-11-08.