Llama 3 70B ShiningValiant2: benchmark results
Valiant Labs' Shining Valiant 2 fine-tune of Llama 3 70B Instruct for science, engineering, and structured technical reasoning. Provider: Meta. Released 2024-04-20. Access: Open.
Aggregate Rating 1568 ± 44, rank #1430 of 10916 rated models, from 14 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open Chinese LLM - GSM8K | 62.62 | Accuracy (%) | 87.6 |
| Open LLM Leaderboard - BBH | 63.38 | Score | 87.3 |
| Open LLM Leaderboard - MMLU-Pro | 48.98 | Score | 87 |
| Open Chinese LLM - CMMLU | 68.74 | Accuracy (%) | 81.4 |
| Open Chinese LLM Leaderboard | 65.3 | Official Nonzero-Task Mean (%) | 81.4 |
| Open LLM Leaderboard - GPQA | 33.05 | Score | 80.7 |
| Open Chinese LLM - C-Eval Semantic | 86.04 | Accuracy (%) | 78.4 |
| Open Chinese LLM - ARC Challenge | 57.85 | Accuracy (%) | 77.4 |
| Open Chinese LLM - WinoGrande | 65.67 | Accuracy (%) | 74.1 |
| Open LLM Leaderboard - IFEval | 61.22 | Score | 73.4 |
| Open LLM Leaderboard - MuSR | 43.26 | Score | 72.7 |
| Open LLM Leaderboard - MATH Level 5 | 20.77 | Score | 72.2 |
Interactive version: theaggregate.ai/model?slug=llama-3-70b-shiningvaliant2 · How It Works · Data refreshed daily, snapshot 2026-11-08.