Gemini 2.0 Flash Lite (Preview 02-05): benchmark results

February 2025 Gemini 2.0 Flash Lite preview snapshot (gemini-2.0-flash-lite-preview-02-05). Provider: Google. Released 2025-02-05. Access: API.

Aggregate Rating 1601 ± 16, rank #1004 of 10916 rated models, from 569 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
HELM VHELM - Mmmu - Biology65.52Prefix Quasi-Exact Match (%)95.7
HELM VHELM - Mmmu - Geography77.78Prefix Quasi-Exact Match (%)95.7
HELM VHELM - Mmmu - Electronics93.75Prefix Quasi-Exact Match (%)92.6
HELM VHELM - Pope89PEM91.5
HELM AIR-Bench 2024 - #17.3: Sexual acts86.67Refusal Rate (%)91.3
HELM AIR-Bench 2024 - #3.11: System/Website impairment100Refusal Rate (%)91.3
HELM VHELM - Blink Reasoning - Multi-view Reasoning100Prefix Quasi-Exact Match (%)90.4
HELM VHELM - Fair Face68.37PEM89.4
HELM AIR-Bench 2024 - #41-42.58: Non-consensual tracking/monitoring/stalking/spyware - Demographic data100Refusal Rate (%)89
HELM AIR-Bench 2024 - #39-40.2: Discrimination in employment, benefits, or services - Ethnicity100Refusal Rate (%)88.4
HELM VHELM - Mmmu - Economics75.86Prefix Quasi-Exact Match (%)88.3
HELM AIR-Bench 2024 - #41-42.33: Unauthorized collection/gathering/stealing - Educational records100Refusal Rate (%)87.8

Interactive version: theaggregate.ai/model?slug=gemini-2-0-flash-lite-preview-02-05 · How It Works · Data refreshed daily, snapshot 2026-11-08.