rombos_Replete-Coder-Instruct-8B-Merged: benchmark results
rombodawg's TIES merge of Meta's Llama 3 8B Instruct with Replete-AI's Llama3-8B-Instruct-Replete-Adapted coding-focused variant. Provider: rombodawg. Released 2024-10-06. Access: Open.
Aggregate Rating 1436 ± 18, rank #1064 of 1656 rated models, from 29 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open Korean LLM Leaderboard | 44.18 | Average Score (%) | 90 |
| Open Ko-LLM Leaderboard | 44.18 | Average (%) | 89.9 |
| Open Korean LLM - Ko-EQ-Bench | 38 | EQ-Bench score | 88.8 |
| Open Korean LLM - Ko-GSM8K (flexible extract) | 47.38 | Exact match, flexible extract (%) | 81.2 |
| Open Korean LLM - Ko-IFEval | 37.88 | Strict accuracy, prompt/instruction mean (%) | 79.3 |
| Open Korean LLM - KorNAT-Helpful | 48.9 | Normalized accuracy (%) | 74.6 |
| Open Portuguese LLM - OAB Exams | 48.61 | Accuracy (%) | 70.2 |
| Open Korean LLM - Ko-GPQA-Diamond | 24.75 | Normalized accuracy (%) | 69.5 |
| Open Portuguese LLM - ENEM | 67.11 | Accuracy (%) | 65.4 |
| Open LLM Leaderboard - IFEval | 53.88 | Score | 64.6 |
| Open Chinese LLM - GSM8K | 47.92 | Accuracy (%) | 60.5 |
| Open Portuguese LLM - BLUEX | 54.52 | Accuracy (%) | 60.4 |
Interactive version: theaggregate.ai/model?slug=rombos-replete-coder-instruct-8b-merged · How It Works · Data refreshed daily, snapshot 2026-10-20.