Mixtral_34Bx2_MoE_60B Benchmark Results

Mixtral_34Bx2_MoE_60B — benchmark results

cloudyu's 2x34B mixture-of-experts merging Yi-34B-based Bagel-DPO and SUS-Chat for English and Chinese; 'Mixtral' in name only. Provider: Mistral. Released 2024-01-05. Access: Open.

Unified ELO 1494 ± 20, rank #824 of 1776 rated models, from 14 benchmark results.

Strongest benchmark results

Benchmark	Score	Metric	Percentile
Open Chinese LLM Leaderboard	69.41	Average Score (%)	95.5
Open Chinese LLM - WinoGrande	70.8	Accuracy (%)	95.1
Open Chinese LLM - C-Eval Semantic	90.49	Accuracy (%)	94.7
Open Chinese LLM - ARC Challenge	63.99	Accuracy (%)	94.4
Open Chinese LLM - HellaSwag	69.64	Accuracy (%)	92
Open LLM Leaderboard - MuSR	17.78	Score	89.4
Open Chinese LLM - CMMLU	70.88	Accuracy (%)	89
Open LLM Leaderboard - MMLU-Pro	41.85	Score	86.5
Open Chinese LLM - GSM8K	60.12	Accuracy (%)	84.7
Open LLM Leaderboard - GPQA	11.74	Score	83.7
Open Chinese LLM - TruthfulQA MC	59.94	Accuracy (%)	83.4
Open LLM Leaderboard - BBH	41.21	Score	80.4

Interactive version: theaggregate.ai/model?slug=mixtral-34bx2-moe-60b · How the rankings work · Data refreshed daily, snapshot 2026-07-22.