MixtureofMerges-MoE-4x7B-v4 Benchmark Results

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: Other. Released 2024-02-11. Access: Open.

Unified ELO 1410 ± 14, rank #1174 of 1776 rated models, from 14 benchmark results.

Strongest benchmark results

Benchmark	Score	Metric	Percentile
Open Chinese LLM - TruthfulQA MC	61.78	Accuracy (%)	89.3
Open LLM Leaderboard - MuSR	13.89	Score	76.4
Open Chinese LLM - WinoGrande	63.85	Accuracy (%)	61.9
Open Chinese LLM - GSM8K	48.07	Accuracy (%)	61.4
Open LLM Leaderboard - BBH	32.22	Score	60.4
Open Chinese LLM - HellaSwag	60.47	Accuracy (%)	52.5
Open Chinese LLM - ARC Challenge	52.65	Accuracy (%)	50.6
Open Chinese LLM Leaderboard	56.61	Average Score (%)	49.6
Open LLM Leaderboard - GPQA	4.81	Score	41
Open LLM Leaderboard - IFEval	40.3	Score	40.3
Open LLM Leaderboard - MMLU-Pro	22.58	Score	37.5
Open LLM Leaderboard - MATH Level 5	6.34	Score	34.6

Interactive version: theaggregate.ai/model?slug=mixtureofmerges-moe-4x7b-v4 · How the rankings work · Data refreshed daily, snapshot 2026-07-22.