Go
gemini-3.5-flash Benchmark & Insights
Google Gemini API
Updated Jul 18, 2026 All models
Sample size
157 runs
in window
Accuracy
93.8%
consensus match · 160d
Confidence
93%
over 162 runs
Window end
Jul 18, 2026
most recent run
Input price
$1.50/MTok
prompt tokens
Output price
$9.00/MTok
completion tokens
Model insights
- 01 Best full-sample model alongside its flash-lite sibling, leaning "optimistic" (9 under vs 1 over) with mostly mild "low"-vs-"medium" slips.
- 02 One bad day stands out: it said "safe" on 2026-02-25 when consensus was "unsafe".
- 03 Costs six times more than flash-lite for identical accuracy.
Notes
A new stable version of Gemini 3.5 Flash
Recent forecasts