ECE-TW3-JRGL-V5: benchmark results
paloalma's mergekit combination of two Qwen-72B derivatives, MoMo-72B DPO and Alpaca Dragon 72B, in the ECE-TW3-JRGL merge series. Provider: paloalma. Released 2024-04-10. Access: Open.
Aggregate Rating 1613 ± 42, rank #897 of 10916 rated models, from 14 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open Chinese LLM - C-Eval Semantic | 90.82 | Accuracy (%) | 97.3 |
| Open Chinese LLM Leaderboard | 71.22 | Official Nonzero-Task Mean (%) | 97.3 |
| Open Chinese LLM - HellaSwag | 72.92 | Accuracy (%) | 97 |
| Open Chinese LLM - ARC Challenge | 64.16 | Accuracy (%) | 95.6 |
| Open Chinese LLM - GSM8K | 69.37 | Accuracy (%) | 94.7 |
| Open Chinese LLM - WinoGrande | 69.22 | Accuracy (%) | 89.6 |
| Open LLM Leaderboard - MuSR | 46.21 | Score | 88.6 |
| Open Chinese LLM - CMMLU | 70.8 | Accuracy (%) | 88.5 |
| Open Chinese LLM - TruthfulQA MC | 61.27 | Accuracy (%) | 88.2 |
| Open LLM Leaderboard - MMLU-Pro | 46.48 | Score | 85.4 |
| Open LLM Leaderboard - GPQA | 34.14 | Score | 84.9 |
| Open LLM Leaderboard - BBH | 60.25 | Score | 82.2 |
Interactive version: theaggregate.ai/model?slug=ece-tw3-jrgl-v5 · How It Works · Data refreshed daily, snapshot 2026-11-08.