Gemini 2.5 Flash Preview 05-20 (thinking) vs Google: Gemma 4 31B
Compare Gemini 2.5 Flash Preview 05-20 (thinking) and Google: Gemma 4 31B, both from Google AI, context windows of 1.0M vs 262K, tested across 8 shared challenges.
Our Verdict
Google: Gemma 4 31B
Gemini 2.5 Flash Preview 05-20 (thinking)Runner-up
No community votes yet. On paper, Google: Gemma 4 31B has the edge — bigger model tier, newer.
Google: Gemma 4 31B is 8.8x cheaper per token — worth considering if cost matters.
Too close to call
API pricing
Cost per 1M tokens
Gemini 2.5 Flash Preview 05-20 (thinking)
Input
$0.15
Output
$3.50
Google: Gemma 4 31B
Input
$0.14
1.1× cheaper
Output
$0.40
8.8× cheaper
Google: Gemma 4 31B is cheaper on both — 1.1× input, 8.8× output
Writing DNA
Style Comparison
Similarity
56%
Google: Gemma 4 31B uses 37.0x more hedging
Gemini 2.5 Flash Preview 05-20 (thinking)
Google: Gemma 4 31B
70%Vocabulary57%
11wSentence Length18w
0.00Hedging0.37
4.2Bold7.1
7.3Lists4.0
0.00Emoji0.24
0.00Headings0.94
0.00Transitions0.16
Based on 3 + 23 text responses
Research
What we learned reading every model
FAQ
