Llemma-7B: benchmark results
Provider: Other. Access: Open.
Unified ELO 1369 ± 17, rank #2358 of 2928 rated models, from 11 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| SWE-Arena | 1002 | Elo Score | 80.6 |
| Open LLM Leaderboard v1 - GSM8K | 33.28 | Accuracy (%) (5-shot) | 49.4 |
| Hallucinations Leaderboard | 41.9 | Average Task Score (score) | 44.7 |
| TheoremQA | 17.2 | Score (%) | 36.4 |
| Open LLM Leaderboard v1 - MMLU | 47.87 | Accuracy (%) (5-shot) | 26.9 |
| Open LLM Leaderboard v1 - ARC Challenge | 46.16 | Normalized accuracy (%) (25-shot) | 21.4 |
| Open LLM Leaderboard v1 - HellaSwag | 62.98 | Normalized accuracy (%) (10-shot) | 16.4 |
| Open LLM Leaderboard v1 - WinoGrande | 63.3 | Accuracy (%) (5-shot) | 16.1 |
| Science Leaderboard | 28 | Average Accuracy (%) | 12.1 |
| BiGGen-Bench | 2.27 | Average Score (1-5) | 10.8 |
| Open LLM Leaderboard v1 - TruthfulQA MC2 | 38.88 | MC2 (%) (0-shot) | 9.4 |
Interactive version: theaggregate.ai/model?slug=llemma-7b · How It Works · Data refreshed daily, snapshot 2026-09-23.