Polarison

DeepSeek · open weights

DeepSeek V4 Pro 0813 API pricing by provider

DeepSeek V4 Pro 0813 is available from 18 providers. The cheapest is Ionstream at $0.96 input / $2.88 output per 1M tokens. DeepSeek’s own API charges $1.32 / $3.96, 27% more than the cheapest host. The priciest host, Venice, charges 1.7x the cheapest.

Prices from the OpenRouter API, retrieved September 15, 2026. Weights: deepseek-ai/DeepSeek-V4-Pro-0813.

ProviderPrecisionInput / 1MOutput / 1MDocument Q&A / monthContextMax output
Ionstream$0.96$2.88$470.401.05M944K
Novitafp8$0.99$2.97$485.101.05M393K
GMICloudfp8$1.056$3.168$517.441.05M944K
DeepInfrafp8$1.30$2.60$598.001.05M16K
NextBitfp8$1.122$3.366$549.781.05M944K
Alibaba$1.122$3.366$549.781M393K
CoreWeavefp8$1.31$3.96$642.801.05M944K
Baidufp8$1.32$3.96$646.801.05M393K
Sail Researchfp4$1.32$3.96$646.801.05M384K
BaseTenfp4$1.32$3.96$646.801.05M262K
Parasailfp8$1.32$3.96$646.801.05M944K
Together$1.32$3.96$646.801.05M944K
DigitalOcean$1.32$3.96$646.801.05M944K
SiliconFlowfp8$1.32$3.96$646.801.05M393K
Fireworks$1.32$3.96$646.801.05M944K
DeepSeekcreator$1.32$3.96$646.801.05M384K
Phala$1.45$4.36$710.801.05M393K
Venice$1.65$4.95$808.501M33K

Sorted by blended price (3 input : 1 output). Document Q&A assumes 50,000 questions a month with 8,000 input and 600 output tokens each. Precision is reported by the provider; “—” means not disclosed.

Compare with DeepSeek’s API page

See benchmarks, cached-input pricing and comparisons for DeepSeek V4 Pro on its model page.

Frequently asked questions

What is the cheapest DeepSeek V4 Pro 0813 API provider?
Ionstream is the cheapest healthy provider we track at $0.96 per 1M input tokens and $2.88 per 1M output tokens.
How much does DeepSeek V4 Pro 0813 cost?
Across 18 providers, input prices range from $0.96 to $1.65 per 1M tokens and output prices from $2.60 to $4.95.
What is DeepSeek V4 Pro 0813’s context window?
DeepSeek V4 Pro 0813 supports up to 1.05M tokens, but some providers serve a smaller context window — check the table before choosing.

Other open-weight models