Llama 2 22B Blocktriangular [bfloat16] [rev: 7adbaa5b8e12]: benchmark results
Provider: chargoddard. Access: Open.
Aggregate Rating 1349 ± 35, rank #7739 of 10916 rated models, from 6 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open LLM Leaderboard v1 - HellaSwag | 82.69 | Normalized accuracy (%) (10-shot) | 53.1 |
| Open LLM Leaderboard v1 - WinoGrande | 75.93 | Accuracy (%) (5-shot) | 44.9 |
| Open LLM Leaderboard v1 - ARC Challenge | 58.28 | Normalized accuracy (%) (25-shot) | 39.4 |
| Open LLM Leaderboard v1 - MMLU | 54.53 | Accuracy (%) (5-shot) | 34.1 |
| Open LLM Leaderboard v1 - GSM8K | 11.22 | Accuracy (%) (5-shot) | 32.5 |
| Open LLM Leaderboard v1 - TruthfulQA MC2 | 39.23 | MC2 (%) (0-shot) | 10.5 |
Interactive version: theaggregate.ai/model?slug=llama-2-22b-blocktriangular-bfloat16-rev-7adbaa5b8e12 · How It Works · Data refreshed daily, snapshot 2026-11-08.