SwallowMaid-8B-L3-SPPO-abliterated: benchmark results
Provider: Other. Access: Open.
Unified ELO 1491 ± 17, rank #1167 of 2656 rated models, from 15 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open Korean LLM - Ko-GSM8K (flexible extract) | 57.77 | Exact match, flexible extract (%) | 92 |
| Open Korean LLM - Ko-IFEval | 42.59 | Strict accuracy, prompt/instruction mean (%) | 86.4 |
| Open Korean LLM - Ko-EQ-Bench | 31.86 | EQ-Bench score | 81 |
| Open Portuguese LLM - ENEM | 71.17 | Accuracy (%) | 79.3 |
| Open Korean LLM Leaderboard | 40.87 | Average Score (%) | 79 |
| Open Ko-LLM Leaderboard | 40.87 | Average (%) | 78.9 |
| Open Portuguese LLM - OAB Exams | 49.98 | Accuracy (%) | 75.4 |
| Open Portuguese LLM - BLUEX | 58.83 | Accuracy (%) | 74.9 |
| Open Portuguese LLM - ASSIN2 RTE | 91.28 | Macro F1 (%) | 65.8 |
| Open Korean LLM - Ko-GPQA-Diamond | 24.24 | Normalized accuracy (%) | 62.4 |
| Open Korean LLM - Ko-Winogrande | 60.85 | Accuracy (%) | 58 |
| Open Korean LLM - KorNAT-Helpful | 47.6 | Normalized accuracy (%) | 57.8 |
Interactive version: theaggregate.ai/model?slug=swallowmaid-8b-l3-sppo-abliterated · How It Works · Data refreshed daily, snapshot 2026-09-19.