Gemini 2.5 Flash Preview (thinking) vs GPT-4.1
Compare Gemini 2.5 Flash Preview (thinking) by Google AI against GPT-4.1 by OpenAI, in 8 community votes, gpt-4.1 wins 71% of head-to-head duels, context windows of 1.0M vs 1.0M, tested across 20 shared challenges.
Reasoning: GPT-4.1 wins 67% of votes
Web Design: Gemini 2.5 Flash Preview (thinking) and GPT-4.1 are tied
Image Generation: GPT-4.1 wins 100% of votes
Our Verdict
GPT-4.1Winner
Gemini 2.5 Flash Preview (thinking)Runner-up
Pick GPT-4.1. In 8 blind votes, GPT-4.1 wins 71% of the time. That's not luck.
GPT-4.1 particularly excels in Image Generation, Reasoning.
Clear winner
API pricing
Cost per 1M tokens
Gemini 2.5 Flash Preview (thinking)
Input
$0.17
11× cheaper
Output
$3.50
2.3× cheaper
GPT-4.1
Input
$2.00
Output
$8.00
Gemini 2.5 Flash Preview (thinking) is cheaper on both — 11× input, 2.3× output
Writing DNA
Style Comparison
Similarity
47%
GPT-4.1 uses 23.7x more emoji
Gemini 2.5 Flash Preview (thinking)
GPT-4.1
53%Vocabulary58%
14wSentence Length19w
0.38Hedging0.38
4.4Bold9.0
4.1Lists6.1
0.00Emoji0.24
0.00Headings1.01
0.24Transitions0.07
Based on 7 + 25 text responses
Research
What we learned reading every model
FAQ
