NeuralBeagle14-7B: benchmark results
mlabonne's DPO fine-tune of his Beagle14-7B Mistral merge, briefly the top-ranked 7B on the Open LLM Leaderboard in January 2024. Provider: mlabonne. Released 2024-01-15. Access: Open.
Aggregate Rating 1480 ± 27, rank #850 of 1656 rated models, from 43 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open LLM Leaderboard v1 - ARC Challenge | 72.95 | Normalized accuracy (%) (25-shot) | 94.5 |
| EuroEval German NLU - Sb10K | 59.6 | Matthews correlation coefficient (%) | 93.1 |
| Open LLM Leaderboard v1 - GSM8K | 70.28 | Accuracy (%) (5-shot) | 92.3 |
| Open LLM Leaderboard v1 - HellaSwag | 88.34 | Normalized accuracy (%) (10-shot) | 91.1 |
| Open LLM Leaderboard v1 - TruthfulQA MC2 | 69.93 | MC2 (%) (0-shot) | 89.7 |
| Open LLM Leaderboard v1 - WinoGrande | 82.4 | Accuracy (%) (5-shot) | 83.9 |
| YALL Nous Leaderboard | 56.9 | Average | 75.9 |
| EuroEval Dutch NLU - CoNLL NL | 63.53 | Micro F1 excluding MISC (%) | 73.7 |
| EuroEval German NLU - GermEval | 64.81 | Micro F1 excluding MISC (%) | 73 |
| Open LLM Leaderboard v1 - MMLU | 64.55 | Accuracy (%) (5-shot) | 72.4 |
| Open LLM Leaderboard - MuSR | 43.19 | Score | 72 |
| EuroEval English Common Sense Reasoning | 71.96 | Common Sense Reasoning Average Score (%) | 69.7 |
Interactive version: theaggregate.ai/model?slug=neuralbeagle14-7b · How It Works · Data refreshed daily, snapshot 2026-10-20.