ChatGPT-4o (March 2025) vs DeepSeek R1
Compare ChatGPT-4o (March 2025) by OpenAI against DeepSeek R1 by DeepSeek, context windows of 128K vs 128K, tested across 32 shared challenges.
Our Verdict
ChatGPT-4o (March 2025)
DeepSeek R1Runner-up
No community votes yet. On paper, ChatGPT-4o (March 2025) has the edge — bigger model tier, newer.
DeepSeek R1 is 6.8x cheaper per token — worth considering if cost matters.
Too close to call
API pricing
Cost per 1M tokens
ChatGPT-4o (March 2025)
Input
$5.00
Output
$15.00
DeepSeek R1
Input
$0.55
9.1× cheaper
Output
$2.19
6.8× cheaper
DeepSeek R1 is cheaper on both — 9.1× input, 6.8× output
Writing DNA
Style Comparison
Similarity
45%
ChatGPT-4o (March 2025) uses 21.9x more emoji
ChatGPT-4o (March 2025)
DeepSeek R1
55%Vocabulary62%
24wSentence Length15w
0.52Hedging0.52
3.3Bold8.0
6.0Lists5.7
0.94Emoji0.04
0.63Headings0.62
0.05Transitions0.28
Based on 15 + 22 text responses
Research
What we learned reading every model
FAQ
