deepseek Open weights

DeepSeek V3 pricing & benchmarks

Released Dec 26, 2024 · served by 3 providers

Input / 1M tokens$0.200
Output / 1M tokens$0.800
Context window164K
Cached input
1M in + 300K out$0.44

What DeepSeek V3 is

DeepSeek-V3 is the latest model from the DeepSeek team, building upon the instruction following and coding abilities of the previous versions. Pre-trained on nearly 15 trillion tokens, the reported evaluations...

Providers and prices for DeepSeek V3

The same model costs different amounts depending on who serves it. Prices are per 1M tokens, cheapest first. Uptime and throughput are OpenRouter's rolling measurements, not vendor claims.

ProviderInputOutputContext QuantUptime 24hThroughput
StreamLakecheapest $0.200 $0.800 128K 99.4%
DeepInfra $0.320 $0.890 164K fp4 92.5%
Novita $0.400 $1.30 64K fp8 99.0%

DeepSeek V3 benchmark results

Design Arena head-to-head results, as reported through the OpenRouter model API.

CategoryArenaRankEloWin rate
3d models #64 1144 50.7%
uicomponent models #70 1131 52.8%
svg models #72 1026 38.8%
website models #76 1145 48.5%
dataviz models #77 1127 51.4%
codecategories models #78 1139 48.5%
gamedev models #80 1110 43.9%

FAQ

How much does the DeepSeek V3 API cost?

$0.200 per 1M input tokens and $0.800 per 1M output tokens. A typical workload of 1M input + 300K output tokens costs about $0.44.

Is DeepSeek V3 free?

No. It is a paid model starting at $0.200 per 1M input tokens, though some providers offer trial credits.

Which provider is cheapest for DeepSeek V3?

StreamLake at $0.200 input / $0.800 output per 1M tokens.

Can I self-host DeepSeek V3?

Yes — weights are published as deepseek-ai/DeepSeek-V3 on Hugging Face, so you can run it on your own hardware.