Mixtral Instruct 8x7B — benchmark results

Provider: Mistral. Released 2023-12-11. Access: Open.

Unified ELO 1353 ± 25, rank #1430 of 1839 rated models, from 29 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
HELM EWoK - Physical Dynamics89.33EM82.1
HELM EWoK - Material Dynamics82.37EM50
HELM EWoK - Material Properties85.08EM35.7
HELM EWoK - Physical Interactions72.32EM35.7
HELM EWoK - Spatial Relations61.84EM35.7
HELM EWoK - Agent Properties79.46EM28.6
HELM EWoK - Physical Relations70.24EM28.6
HELM EWoK - Social Interactions95.14EM28.6
HELM Safety BBQ85.7BBQ accuracy (%)23.3
HELM Safety XSTest93.1LM Evaluated Safety score (%)23.3
HELM ThaiExam - TPAT-135.34EM19.5
HELM Finance - BANKING7759.8EM19

Interactive version: theaggregate.ai/model?slug=mixtral-instruct-8x7b · How It Works · Data refreshed daily, snapshot 2026-08-05.