Gemini 1.5 Pro (001, BLOCK_NONE safety): benchmark results

Provider: Google. Access: API.

Aggregate Rating 1619 ± 22, rank #844 of 10916 rated models, from 506 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
HELM AIR-Bench 2024 - #41-42.49: Unauthorized inference/synthesis - Demographic data94.44Refusal Rate (%)99.4
HELM AIR-Bench 2024 - #5.18: Life support100Refusal Rate (%)99.4
HELM AIR-Bench 2024 - #4.6: Housing eligibility73.33Refusal Rate (%)98.8
HELM AIR-Bench 2024 - #4.8: Migration eligibility56.67Refusal Rate (%)98.8
HELM AIR-Bench 2024 - #4: Automated Decision-Making78Refusal Rate (%)98.8
HELM AIR-Bench 2024 - #5.17: Water treatment facilities96.67Refusal Rate (%)98.8
HELM AIR-Bench 2024 - Operational Misuses69.74Refusal Rate (%)98.8
HELM AIR-Bench 2024 - #4.4: Employment73.33Refusal Rate (%)98.3
HELM AIR-Bench 2024 - #5.20: Emergency services60Refusal Rate (%)98.3
HELM AIR-Bench 2024 - #4.1: Financing eligibility/Creditworthiness83.33Refusal Rate (%)97.7
HELM AIR-Bench 2024 - #41-42.13: Unauthorized disclosure - Demographic data100Refusal Rate (%)97.7
HELM AIR-Bench 2024 - #5: Autonomous Unsafe Operation of Systems81.52Refusal Rate (%)97.7

Interactive version: theaggregate.ai/model?slug=gemini-1-5-pro-001-block-none-safety · How It Works · Data refreshed daily, snapshot 2026-11-08.