Llama 3.1 8B ShiningValiant2: benchmark results
Valiant Labs' science-and-engineering fine-tune of Llama 3.1 8B Instruct, emphasizing technical knowledge and structured reasoning. Provider: Valiant Labs. Released 2024-08-06. Access: Open.
Aggregate Rating 1378 ± 19, rank #1306 of 1661 rated models, from 21 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open Korean LLM - Ko-IFEval | 37.29 | Strict accuracy, prompt/instruction mean (%) | 77.8 |
| Open LLM Leaderboard - IFEval | 64.96 | Score | 77.2 |
| Open Portuguese LLM - FaQuAD NLI | 75.95 | Macro F1 (%) | 75.1 |
| Open LLM Leaderboard - GPQA | 31.04 | Score | 67.9 |
| Open Korean LLM - Ko-GSM8K (flexible extract) | 29.64 | Exact match, flexible extract (%) | 64.1 |
| Open Portuguese LLM - OAB Exams | 44.33 | Accuracy (%) | 57.9 |
| Open Ko-LLM Leaderboard | 36.45 | Average (%) | 57.4 |
| Open Korean LLM Leaderboard | 36.45 | Average Score (%) | 57.4 |
| Open Portuguese LLM - ASSIN2 RTE | 90.29 | Macro F1 (%) | 56 |
| Open Korean LLM - Ko-EQ-Bench | 7.18 | EQ-Bench score | 53.3 |
| Open Portuguese LLM - ENEM | 61.58 | Accuracy (%) | 49.8 |
| Open LLM Leaderboard - MMLU-Pro | 33.82 | Score | 48.3 |
Interactive version: theaggregate.ai/model?slug=llama-3-1-8b-shiningvaliant2 · How It Works · Data refreshed daily, snapshot 2026-10-23.