GitHub[208K]Try OpenCode

Compare AI models: DeepSeek V4 Flash Vision Exp vs DeepSeek V4 Pro

  1. DeepSeek V4 Flash Vision ExpDeepSeek
  2. DeepSeek V4 ProDeepSeek
ReasoningCodingCost efficiencyContext windowMultimodalTool use
Normalized model capability scores
ModelReasoningCodingCost efficiencyContext windowMultimodalTool use
DeepSeek V4 Flash Vision Exp27/10079/10060/10080/10033/1000/100
DeepSeek V4 Pro100/100100/10037/10080/1000/10068/100
Overview
Author
Context length
1M
1M
Reasoning
True
True
Input modalities
Text, Image
Text
Output modalities
Text
Text
Providers
Pricing
Input
$0.30/ 1M
$1.74/ 1M
Output
$1.20/ 1M
$3.48/ 1M
Cached input
$0.006/ 1M
$0.20/ 1M
MomentumLast 2 mo
Unique users
335K
911K
Completed sessions
979,932
1,345,351
Token share
0.92%
1.7%
Tokens
8.8T+100%
16T-45%
AUG 11OCT 5
AUG 11OCT 5
RetentionWeek 1
Returning users
62.1%24K user-weeks
55.7%81K user-weeks

Related comparisons. Other model pairs to check.