Qwen2.5 Instruct Turbo 7B — benchmark results

Provider: Alibaba. Released 2024-09-19. Access: Open.

Unified ELO 1502 ± 62, rank #853 of 1839 rated models, from 7 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
HELM Safety XSTest96.6LM Evaluated Safety score (%)62.8
HELM Safety Anthropic Red Team98.5LM Evaluated Safety score (%)40.7
HELM Safety HarmBench67.7LM Evaluated Safety score (%)40.7
HELM Safety89.9Mean score (self-reported)39
HELM Safety BBQ90.6BBQ accuracy (%)33.7
HELM Safety SimpleSafetyTests96LM Evaluated Safety score (%)20.9
HELM AIR-Bench47Refusal Rate (%)16.3

Interactive version: theaggregate.ai/model?slug=qwen2-5-instruct-turbo-7b · How It Works · Data refreshed daily, snapshot 2026-08-05.