MixtureofMerges-MoE-4x7B-v5: benchmark results
jsfs11's LazyMergekit 4x7B MoE routing OmniBeagleSquaredMBX, AlphaMonarch, Neural4gsm8k and NeutrixOmnibe-DPO experts by task. Provider: jsfs11. Released 2024-02-25. Access: Open.
Aggregate Rating 1468 ± 23, rank #896 of 1656 rated models, from 25 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open LLM Leaderboard v1 - ARC Challenge | 73.89 | Normalized accuracy (%) (25-shot) | 98.5 |
| Open LLM Leaderboard v1 - WinoGrande | 85.08 | Accuracy (%) (5-shot) | 97.1 |
| Open LLM Leaderboard v1 - HellaSwag | 89 | Normalized accuracy (%) (10-shot) | 95.8 |
| Open LLM Leaderboard v1 - TruthfulQA MC2 | 73.73 | MC2 (%) (0-shot) | 94.5 |
| Open Chinese LLM - TruthfulQA MC | 62.58 | Accuracy (%) | 91.1 |
| Open LLM Leaderboard v1 - GSM8K | 69.75 | Accuracy (%) (5-shot) | 90.3 |
| Open Portuguese LLM - FaQuAD NLI | 78.23 | Macro F1 (%) | 83 |
| Open Portuguese LLM - ASSIN2 RTE | 92.36 | Macro F1 (%) | 78.2 |
| Open LLM Leaderboard v1 - MMLU | 64.69 | Accuracy (%) (5-shot) | 75 |
| Open LLM Leaderboard - MuSR | 43.05 | Score | 70.7 |
| Open Chinese LLM - WinoGrande | 64.09 | Accuracy (%) | 64.1 |
| Open Portuguese LLM - BLUEX | 54.8 | Accuracy (%) | 61.4 |
Interactive version: theaggregate.ai/model?slug=mixtureofmerges-moe-4x7b-v5 · How It Works · Data refreshed daily, snapshot 2026-10-20.