Llemma-MetaMath-7B: benchmark results
Provider: Other. Access: Open.
Aggregate Rating 1332 ± 30, rank #8116 of 10916 rated models, from 18 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open LLM Leaderboard v1 - GSM8K | 60.96 | Accuracy (%) (5-shot) | 73.3 |
| Open LLM Leaderboard v1 - MMLU | 47.66 | Accuracy (%) (5-shot) | 26.6 |
| GR-Ben - Chemistry | 22.8 | F1 (%) of first-error identification: the harmonic mean of t | 23.8 |
| GR-Ben - Computer Science | 18.1 | F1 (%) of first-error identification: the harmonic mean of t | 23.8 |
| GR-Ben - Inductive Reasoning | 18.1 | F1 (%) of first-error identification: the harmonic mean of t | 23.8 |
| Open LLM Leaderboard v1 - ARC Challenge | 46.5 | Normalized accuracy (%) (25-shot) | 21.6 |
| MathBench | 24.9 | Average Score | 21.2 |
| PRMBench | 50.5 | Overall Score | 16 |
| Open LLM Leaderboard v1 - WinoGrande | 62.75 | Accuracy (%) (5-shot) | 15.7 |
| Open LLM Leaderboard v1 - HellaSwag | 61.69 | Normalized accuracy (%) (10-shot) | 15.6 |
| GR-Ben - Abductive Reasoning | 8.3 | F1 (%) of first-error identification: the harmonic mean of t | 14.3 |
| GR-Ben - Deductive Reasoning | 11.4 | F1 (%) of first-error identification: the harmonic mean of t | 14.3 |
Interactive version: theaggregate.ai/model?slug=llemma-metamath-7b · How It Works · Data refreshed daily, snapshot 2026-11-08.