Mistral-Small-Instruct-2409-abliterated: benchmark results
byroneverson's abliterated variant of Mistral Small Instruct 2409 (22B) with refusal behavior removed. Provider: byroneverson. Released 2024-09-23. Access: Open.
Aggregate Rating 1497 ± 18, rank #795 of 1656 rated models, from 29 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open Portuguese LLM - FaQuAD NLI | 86.08 | Macro F1 (%) | 99 |
| Open Korean LLM - Ko-EQ-Bench | 58.61 | EQ-Bench score | 98 |
| Open Korean LLM - Ko-IFEval | 55.05 | Strict accuracy, prompt/instruction mean (%) | 96.5 |
| Open Korean LLM Leaderboard | 48 | Average Score (%) | 95.6 |
| Open Ko-LLM Leaderboard | 48 | Average (%) | 95.5 |
| Open Portuguese LLM - ASSIN2 RTE | 93.41 | Macro F1 (%) | 89.2 |
| Open Portuguese LLM - OAB Exams | 54.21 | Accuracy (%) | 85.6 |
| Open Korean LLM - Ko-GSM8K (flexible extract) | 52.24 | Exact match, flexible extract (%) | 85.3 |
| Open Chinese LLM - GSM8K | 59.51 | Accuracy (%) | 83.4 |
| Open Korean LLM - Ko-Winogrande | 65.11 | Accuracy (%) | 82.7 |
| Open LLM Leaderboard - IFEval | 69.71 | Score | 82.7 |
| Open LLM Leaderboard - GPQA | 33.31 | Score | 81.7 |
Interactive version: theaggregate.ai/model?slug=mistral-small-instruct-2409-abliterated · How It Works · Data refreshed daily, snapshot 2026-10-20.