All models

DeepSeek V4 Pro

deepseek/deepseek-v4-pro

DeepSeek V4 Pro is a large-scale Mixture-of-Experts model from DeepSeek with 1.6T total parameters and 49B activated parameters, supporting a 1M-token context window. It is designed for advanced reasoning, coding,...

Modalities
Text
In / Out price
$0.435 / $0.87 per 1M
Context
1.05M · 384K out
Released
Apr 24, 2026

Providers

Live list prices and uptime across 18 providers serving this model.

ProviderInput /MOutput /MCache read /MContextMax outputUptime 24h
DeepSeek$0.435$0.87$0.0041.05M384K99.89%
Baidufp8$0.625$1.251$0.0521.05M393K96.06%
StreamLakefp8$0.67$1.34$0.0561.02M384K95.49%
GMICloudfp8$0.679$1.357$0.0571.05M92.95%
Ionstreamfp4$1.131$2.262$0.0941.05M393K57.81%
Novitafp8$1.168$2.336$0.0991.05M393K99.13%
Waferfp4$1.20$2.40$0.101.05M131K91.26%
DeepInfrafp4$1.30$2.60$0.101.05M16K99.11%
DigitalOcean$1.392$2.784$0.348262K91.08%
Alibaba$1.416$2.832$0.1181M393K93.06%
SiliconFlowfp8$1.502$3.135$0.1351.05M393K97.71%
Venice$1.65$3.301$0.331M33K85.74%
AtlasCloudfp4$1.68$3.38$0.131.05M393K98.43%
BaseTenfp4$1.74$3.48$0.145262K262K99.33%
Parasailfp8$1.74$3.48$0.101.05M1.05M86.60%
Together$1.74$3.48$0.20512K94.14%
CoreWeavefp8$1.74$3.48$0.141.05M1.05M97.42%
Fireworks$1.74$3.48$0.1451.05M87.14%

Implicit prompt caching available — repeated context bills at the cache read rate.

Capabilities

Tool calling
Reasoning
frequency_penalty
include_reasoning
logit_bias
logprobs
max_tokens
min_p
presence_penalty
reasoning
reasoning_effort
repetition_penalty
response_format
seed
stop
structured_outputs
temperature
tool_choice
top_k
top_logprobs
top_p

Pricing detail

Input
$0.435
Output
$0.87
Cache read
$0.004
Cache write

USD per 1M tokens.