Llama 3.1 8B Cobalt: benchmark results
Valiant Labs' math-instruct fine-tune of Llama 3.1 8B Instruct, trained on synthetic math data generated with Llama 3.1 405B. Provider: Valiant Labs. Released 2024-08-16. Access: Open.
Aggregate Rating 1427 ± 20, rank #1114 of 1654 rated models, from 20 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open Korean LLM - Ko-IFEval | 43.44 | Strict accuracy, prompt/instruction mean (%) | 87.8 |
| Open Portuguese LLM - FaQuAD NLI | 79.33 | Macro F1 (%) | 86.9 |
| Open LLM Leaderboard - IFEval | 71.68 | Score | 85.2 |
| Open Korean LLM Leaderboard | 40.29 | Average Score (%) | 76.3 |
| Open Korean LLM - Ko-EQ-Bench | 26.57 | EQ-Bench score | 73.3 |
| Open Korean LLM - Ko-GPQA-Diamond | 24.75 | Normalized accuracy (%) | 69.5 |
| Open Portuguese LLM - ENEM | 68.02 | Accuracy (%) | 68.3 |
| Open Portuguese LLM - OAB Exams | 46.51 | Accuracy (%) | 64 |
| Open Portuguese LLM - ASSIN2 RTE | 91.03 | Macro F1 (%) | 62.7 |
| Open LLM Leaderboard - MATH Level 5 | 15.33 | Score | 62.5 |
| Open Korean LLM - KorNAT-Helpful | 47.78 | Normalized accuracy (%) | 60.3 |
| Open LLM Leaderboard - GPQA | 30.37 | Score | 60.3 |
Interactive version: theaggregate.ai/model?slug=llama-3-1-8b-cobalt · How It Works · Data refreshed daily, snapshot 2026-10-24.