Llama 2 7B Base: benchmark results
Meta's Llama 2 7B, the base pretrained checkpoint released under the Llama 2 Community License in July 2023; the linked model card provides no further details here. Provider: Meta. Released 2023-07-18. Access: Open.
Aggregate Rating 1271 ± 10, rank #1518 of 1656 rated models, from 277 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| EuroEval Swedish NLU - Swerec | 79.05 | Matthews correlation coefficient (%) | 90.3 |
| MERA Original Paper - ruWorldTree | 54.5 | Accuracy (%; 5-shot, log-likelihood) | 83.3 |
| MMLU-by-task - High School Mathematics | 30.37 | Accuracy (%) | 79.4 |
| MERA Original Paper | 32.7 | Total Score (%; mean of 17 problem-solving and exam tasks, M | 77.8 |
| MERA Original Paper - MathLogicQA | 27.7 | Accuracy (%; 5-shot, log-likelihood) | 77.8 |
| MERA Original Paper - SimpleAr | 83.9 | Exact Match (%; 5-shot, greedy generation) | 77.8 |
| MERA Original Paper - ruMMLU | 45.2 | Accuracy (%; 5-shot, log-likelihood) | 77.8 |
| MERA Original Paper - ruModAr | 36.7 | Exact Match (%; greedy generation) | 77.8 |
| MERA Original Paper - ruMultiAr | 12.4 | Exact Match (%; 5-shot, greedy generation) | 77.8 |
| MERA Original Paper - ruOpenBookQA | 47.5 | Accuracy (%; 5-shot, log-likelihood) | 77.8 |
| MMLU-by-task - Machine Learning | 38.39 | Accuracy (%) | 75.5 |
| MMLU-by-task - College Mathematics | 35 | Accuracy (%) | 75 |
Interactive version: theaggregate.ai/model?slug=llama-2-7b-base · How It Works · Data refreshed daily, snapshot 2026-10-20.