Gemini 2.5 Flash Preview (thinking) vs Gemini 3 Pro Preview
Compare Gemini 2.5 Flash Preview (thinking) and Gemini 3 Pro Preview, both from Google AI, tested across 18 shared challenges.
Our Verdict
Gemini 3 Pro Preview
Gemini 2.5 Flash Preview (thinking)Runner-up
No community votes yet. On paper, Gemini 3 Pro Preview has the edge — bigger model tier, newer.
Too close to call
API pricing
Cost per 1M tokens
Gemini 2.5 Flash Preview (thinking)
Input
$0.17
Output
$3.50
Gemini 3 Pro Preview
Input
—
Output
—
Writing DNA
Style Comparison
Similarity
50%
Gemini 3 Pro Preview uses 89.8x more headings
Gemini 2.5 Flash Preview (thinking)
Gemini 3 Pro Preview
53%Vocabulary58%
14wSentence Length17w
0.38Hedging0.37
4.4Bold7.7
4.1Lists4.8
0.00Emoji0.00
0.00Headings0.90
0.24Transitions0.25
Based on 7 + 23 text responses
Research
What we learned reading every model
FAQ
