CarbonBeagle-11B: benchmark results
vicgalle's linear merge of his NeuralBeagle-11B with the SOLAR-based CarbonVillain-10.7B, briefly the top 11B-13B model on the Open LLM Leaderboard. Provider: vicgalle. Released 2024-01-21. Access: Open.
Aggregate Rating 1460 ± 15, rank #937 of 1656 rated models, from 35 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open LLM Leaderboard v1 - HellaSwag | 88.93 | Normalized accuracy (%) (10-shot) | 95.1 |
| Open Korean LLM - KorNAT-Harmless | 68.97 | Normalized accuracy (%) | 92.6 |
| Open LLM Leaderboard v1 - WinoGrande | 84.06 | Accuracy (%) (5-shot) | 91.8 |
| Open LLM Leaderboard v1 - ARC Challenge | 71.84 | Normalized accuracy (%) (25-shot) | 90.3 |
| Open LLM Leaderboard v1 - TruthfulQA MC2 | 69.43 | MC2 (%) (0-shot) | 89 |
| Open LLM Leaderboard v1 - MMLU | 66.62 | Accuracy (%) (5-shot) | 87.7 |
| Open Korean LLM - Ko-GSM8K (flexible extract) | 53.45 | Exact match, flexible extract (%) | 87 |
| Open Chinese LLM - TruthfulQA MC | 60.79 | Accuracy (%) | 84.3 |
| Open Korean LLM Leaderboard | 41.94 | Average Score (%) | 84.1 |
| Open Ko-LLM Leaderboard | 41.94 | Average (%) | 84 |
| Open LLM Leaderboard v1 - GSM8K | 66.94 | Accuracy (%) (5-shot) | 82.7 |
| Open Korean LLM - KorNAT-Helpful | 49.7 | Normalized accuracy (%) | 81 |
Interactive version: theaggregate.ai/model?slug=carbonbeagle-11b · How It Works · Data refreshed daily, snapshot 2026-10-20.