Nxcode-CQ-7B-orpo: benchmark results
NTQ's Nxcode-CQ-7B-orpo, an ORPO fine-tune of Qwen's CodeQwen1.5-7B on 100,000 samples of high-quality ranking data for code generation. Provider: NTQ. Released 2024-04-24. Access: Open.
Aggregate Rating 1302 ± 19, rank #1477 of 1656 rated models, from 30 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| BigCode Models Leaderboard | 87.2 | HumanEval Python Pass@1 (%) | 99.2 |
| Open Korean LLM - Ko-GPQA-Diamond | 25.25 | Normalized accuracy (%) | 75.9 |
| Open Korean LLM - Ko-IFEval | 34.74 | Strict accuracy, prompt/instruction mean (%) | 71.2 |
| Open Korean LLM - KorNAT-Helpful | 47.75 | Normalized accuracy (%) | 60 |
| BigCodeBench | 39.6 | Pass@1 (%) | 56.4 |
| McEval | 44.7 | AVG (%) | 52.4 |
| Open LLM Leaderboard v1 - GSM8K | 27.07 | Accuracy (%) (5-shot) | 45 |
| Open Korean LLM - KorNAT-Harmless | 62.24 | Normalized accuracy (%) | 40.9 |
| Open Korean LLM - Ko-GSM8K (flexible extract) | 13.19 | Exact match, flexible extract (%) | 40.4 |
| Open LLM Leaderboard - IFEval | 40.07 | Score | 39.9 |
| Open Portuguese LLM - ASSIN2 RTE | 86.68 | Macro F1 (%) | 38.4 |
| Open LLM Leaderboard - MuSR | 39.4 | Score | 37.7 |
Interactive version: theaggregate.ai/model?slug=nxcode-cq-7b-orpo · How It Works · Data refreshed daily, snapshot 2026-10-20.