MixtureofMerges-MoE-4x7B-v4: benchmark results
jsfs11's LazyMergekit 4x7B MoE combining MBX-v3, NeuTrixOmniBe-remix, WestSeverus-DPO and supermario_v4 Mistral merges as task experts. Provider: jsfs11. Released 2024-02-11. Access: Open.
Aggregate Rating 1467 ± 23, rank #902 of 1656 rated models, from 25 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open LLM Leaderboard v1 - GSM8K | 71.34 | Accuracy (%) (5-shot) | 95.7 |
| Open LLM Leaderboard v1 - TruthfulQA MC2 | 75.3 | MC2 (%) (0-shot) | 95.5 |
| Open LLM Leaderboard v1 - WinoGrande | 84.85 | Accuracy (%) (5-shot) | 95.5 |
| Open LLM Leaderboard v1 - HellaSwag | 88.85 | Normalized accuracy (%) (10-shot) | 94.4 |
| Open LLM Leaderboard v1 - ARC Challenge | 72.53 | Normalized accuracy (%) (25-shot) | 92.4 |
| Open Chinese LLM - TruthfulQA MC | 61.78 | Accuracy (%) | 89.1 |
| Open Portuguese LLM - FaQuAD NLI | 78.49 | Macro F1 (%) | 84.3 |
| Open LLM Leaderboard - MuSR | 43.86 | Score | 77.2 |
| Open Portuguese LLM - ASSIN2 RTE | 92.24 | Macro F1 (%) | 76.6 |
| Open LLM Leaderboard v1 - MMLU | 64.53 | Accuracy (%) (5-shot) | 72.1 |
| Open Chinese LLM - WinoGrande | 63.85 | Accuracy (%) | 61.7 |
| Open Chinese LLM - GSM8K | 48.07 | Accuracy (%) | 61.5 |
Interactive version: theaggregate.ai/model?slug=mixtureofmerges-moe-4x7b-v4 · How It Works · Data refreshed daily, snapshot 2026-10-20.