Go

gemini-3.1-flash-lite-preview Benchmark & Insights

Google Gemini API
Updated May 25, 2026 All models
Sample size
64 runs
in window
Accuracy
93.8%
consensus match · 64d
Confidence
97%
over 64 runs
Window end
May 25, 2026
most recent run
Input price
$0.25/MTok
prompt tokens
Output price
$1.50/MTok
completion tokens
Model insights
  • 01 Near-flawless on its 64-day preview window, with all four misses being gentle "low" calls against a "medium" consensus and a perfect "safe" record.
  • 02 It was retired in May in favor of the stable gemini-3.1-flash-lite, which holds nearly the same accuracy over the full 155 days.
Notes

Dropping a preview status on May 25, 2026, going to a new stage as gemini-3.1-flash-lite

Recent forecasts