Thinking Levels · Google
Last updated , from board data generated at 17:52 UTC.
This page compares Gemini 3.7 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 3.7 Flash ranks #57 of 87 models through its best setting. The other ranks on this page are configuration ranks: one setting's place among all 151 configurations.
High thinking moves Gemini 3.7 Flash's Writing Elo only a little, and the bill a lot more. High and the medium thinking you get with no flag land 60 Elo apart, with overlapping ranges. High spends several times the reasoning tokens, costs 2.3 times as much per script and takes about 2.2 minutes instead of 0.9.
Stay on medium thinking. The score that moves most is length: at high thinking, the scripts land a little closer to their target, though they still run long at both settings. That isn't worth 2.3 times the price and the longer wait to me, and medium thinking already does the job for $0.038 a script. If you want a better Google script, Gemini 2.5 Pro (dynamic thinking) ranks above both settings for $0.067.
Each point is one effort setting of Gemini 3.7 Flash, joined from the lowest setting to the highest. Higher and further left is better: a higher score for less money or less time.
On a small screen, Table lists the same numbers as rows.
| Setting | Writing Elo | Step | Writing Score | Cost / Script | Time / Script | Reasoning Tokens | Configuration Rank |
|---|---|---|---|---|---|---|---|
| medium thinking (provider default) | 1014 ±48 | - | 69.4 | $0.038 | 0.9 min | 3,971 | #118 |
| high | 1074 ±54 | +60 | 70.1 | $0.086 | 2.2 min | 16,839 | #112 |
The 95% Writing Elo ranges overlap: Gemini 3.7 Flash (medium thinking) runs from 967 to 1063, and Gemini 3.7 Flash (high thinking) from 1018 to 1126.
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.