Towards AITowards AIToneBench

Thinking Levels · Google

Gemini 3.8 Flash: Writing Quality by Reasoning Effort

Last updated , from board data generated at 17:52 UTC.

This page compares Gemini 3.8 Flash at 2 reasoning-effort settings, each tested on the same 10 scripts: the Writing Score (the 0 to 100 consensus of the judges and code), the Writing Elo (a rating from pairwise comparisons of those scores), and the cost and time of each setting. As a model, Gemini 3.8 Flash ranks #56 of 87 models through its best setting. The other ranks on this page are configuration ranks: one setting's place among all 151 configurations.

What More Effort Changes for Gemini 3.8 Flash

High thinking gives Gemini 3.8 Flash a small lead on Writing Elo. High lands 74 Elo above the medium thinking you get with no flag, going from #121 to #111, and the two ranges still overlap. The model really does think more at high, several times the reasoning tokens, and you pay for it: 2.4 times the price per script and about 2.2 minutes instead of 0.9.

Recommended Setting

If you're on Gemini 3.8 Flash, high is a fair bet. The score leans that way, length gets better with it, and the price per script stays low. Just don't expect it to fix the voice. Its tone score is low at both settings, and more thinking doesn't move it.

The bigger thing I noticed across Google's ladders: the older Gemini 2.5 Pro (dynamic thinking) ranks above every Gemini 3 setting on the board, this one included, for $0.067 a script. Its range overlaps this model's at high; overlap alone doesn't settle their order, but it's cheaper and faster for a score in the same range.

Every Setting on the Chart

Each point is one effort setting of Gemini 3.8 Flash, joined from the lowest setting to the highest. Higher and further left is better: a higher score for less money or less time.

How to Use the Chart
  • A hollow marker is the provider default: the level the provider applies when the request sets no effort level.
  • The tabs choose the score on the vertical axis, X axis switches between cost and time, and Table lists the same numbers as rows.
  • Select a point for its numbers and a link to its scorecard.
Score on the vertical axis
View
X axis

On a small screen, Table lists the same numbers as rows.

Both Effort Settings Compared

Score colors:82 and up65 to 81.9under 65
SettingWriting EloStepWriting ScoreCost / ScriptTime / ScriptReasoning TokensConfiguration Rank
medium thinking (provider default)1003 ±46-69.1$0.0360.9 min3,372#121
high1077 ±43+7470.2$0.0872.2 min17,014#111

The 95% Writing Elo ranges overlap: Gemini 3.8 Flash (medium thinking) runs from 952 to 1044, and Gemini 3.8 Flash (high thinking) from 1033 to 1118.

Other Google Models by Effort

Writing Elo comes from comparing every pair of models on every script, which puts these settings on the same scale as the leaderboard. Step is the Writing Elo change from the provider default. Cost is one full script at API list prices, and time is the average for one accepted script, retries included. The thinking-levels page covers the other models, and the Google family page lists every Google model side by side.