DeepSeek V4 Flash 0731 vs Mistral Devstral Medium
Compare DeepSeek V4 Flash 0731 by DeepSeek against Mistral Devstral Medium by Mistral AI, tested across 42 shared challenges.
Our Verdict
DeepSeek V4 Flash 0731
Mistral Devstral MediumRunner-up
No community votes yet. On paper, DeepSeek V4 Flash 0731 has the edge — bigger model tier, newer, bigger context window, major provider backing.
DeepSeek V4 Flash 0731 is 7.1x cheaper per token — worth considering if cost matters.
Slight edge
API pricing
Cost per 1M tokens
DeepSeek V4 Flash 0731
Input
$0.14
2.9× cheaper
Output
$0.28
7.1× cheaper
Mistral Devstral Medium
Input
$0.40
Output
$2.00
DeepSeek V4 Flash 0731 is cheaper on both — 2.9× input, 7.1× output
Writing DNA
Style Comparison
Similarity
54%
Mistral Devstral Medium uses 8.8x more transitions
DeepSeek V4 Flash 0731
Mistral Devstral Medium
52%Vocabulary50%
18wSentence Length17w
0.38Hedging0.56
5.7Bold4.5
3.4Lists5.3
0.00Emoji0.00
0.83Headings0.19
0.04Transitions0.32
Based on 23 + 21 text responses
Research
What we learned reading every model
FAQ
