DeepSeek V4
DeepSeek
· deepseek-v4
Reasoning and Strategy
Production
Lowest Cost, New
Open reasoning model with near-frontier chain-of-thought quality at a fraction of the cost.
Best use cases: Analysis, planning, budget reasoning workloads
8.8
Quality
7
Speed
9.8
Cost Eff.
6.5
Brand Safety
6.5
Control
8.5
Adherence
Input cost
$0.28 / 1M tokens
Output cost
$1.10 / 1M tokens
Generation cost
—
Max context / duration
256K tokens
Aspect ratios
—
Resolutions
—
Input types
Text
Output types
Text
API endpoint
api.deepseek.com
Commercial use
Allowed
Data retention
Provider-dependent
Regions available
Global (check compliance)
API connected
Connected
Last updated
Release date
Documentation