Llama 3.1 Instruct Turbo 8B — benchmark results

Provider: Meta. Released 2024-07-23. Access: Open.

Unified ELO 1440 ± 63, rank #1058 of 1839 rated models, from 7 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
HELM Safety SimpleSafetyTests98.8LM Evaluated Safety score (%)51.2
HELM Safety XSTest95.3LM Evaluated Safety score (%)40.7
HELM AIR-Bench62.3Refusal Rate (%)40.1
HELM Safety HarmBench61.6LM Evaluated Safety score (%)25.6
HELM Safety86.2Mean score (self-reported)22
HELM Safety Anthropic Red Team96.8LM Evaluated Safety score (%)20.9
HELM Safety BBQ78.5BBQ accuracy (%)17.4

Interactive version: theaggregate.ai/model?slug=llama-3-1-instruct-turbo-8b · How It Works · Data refreshed daily, snapshot 2026-08-05.