Llama-3.1-8B-Lexi-Uncensored: benchmark results
Orenguteng's uncensored fine-tune of Llama 3.1 8B Instruct that strips refusals for maximum compliance; superseded by Lexi V2. Provider: Meta. Released 2024-07-26. Access: Open.
Aggregate Rating 1466 ± 40, rank #2966 of 10916 rated models, from 15 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open LLM Leaderboard - IFEval | 77.77 | Score | 93.8 |
| Open Korean LLM - Ko-GSM8K (flexible extract) | 55.5 | Exact match, flexible extract (%) | 89.3 |
| Open Korean LLM Leaderboard | 43.66 | Average Score (%) | 88.6 |
| Open Korean LLM - Ko-EQ-Bench | 30.78 | EQ-Bench score | 79.9 |
| Open Korean LLM - Ko-IFEval | 38.36 | Strict accuracy, prompt/instruction mean (%) | 79.9 |
| Open Korean LLM - KorNAT-CKA | 28.94 | Normalized accuracy (%) | 77.9 |
| Open Korean LLM - Ko-GPQA-Diamond | 25.25 | Normalized accuracy (%) | 75.9 |
| Open LLM Leaderboard - MMLU-Pro | 37.9 | Score | 64.9 |
| Open LLM Leaderboard - MATH Level 5 | 15.71 | Score | 63.2 |
| Open Korean LLM - Ko-Winogrande | 60.14 | Accuracy (%) | 51.6 |
| Open LLM Leaderboard - BBH | 50.57 | Score | 50.8 |
| Open LLM Leaderboard - MuSR | 38.71 | Score | 33.4 |
Interactive version: theaggregate.ai/model?slug=llama-3-1-8b-lexi-uncensored · How It Works · Data refreshed daily, snapshot 2026-11-08.