Gemini 3.5 Flash Lite (High): benchmark results

Provider: Google. Released 2026-07-21. Access: API.

Unified ELO 1654 ± 1, rank #311 of 3081 rated models, from 71 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
Korean CSAT 2026 (Easy Mode) - Chemistry I50Points (out of 50)90.8
Vals AI MortgageTax68.68Accuracy (%)87.6
Korean CSAT 2026 (Easy Mode) - Physics I43Points (out of 50)81
ZeroBench14Score (%)75.4
Vals AI LegalBench84.07Accuracy (%)75.3
Korean CSAT 2026 (Easy Mode) - Life Science I45Points (out of 50)74.5
IOI26.17Score (self-reported)74.2
LiveBench Simplify68.02Score71.9
Korean CSAT 2026 (Easy Mode) - Total422.67Points (out of 450)70.1
Epoch AI - Dtbench83.47Score69.4
Epoch AI - GPQA Diamond83.33Accuracy (%)68.1
Korean CSAT 2026 (Easy Mode) - Society and Culture42Points (out of 50)68

Interactive version: theaggregate.ai/model?slug=gemini-3-5-flash-lite-high · How It Works · Data refreshed daily, snapshot 2026-09-21.