DeepSeek API Providers

Who serves DeepSeek, at what price and speed.

DeepSeek releases its models with open weights, from V3 and R1 to the V4 Flash and Pro generation and V4.1 Flash.

DeepSeek served 14% of the 81.1T DeepSeek tokens that went through a large public LLM router over Nov 2 to Nov 8, 2026. DeepSeek's own API served 14%; 46 providers list the family.

Who serves DeepSeek tokens

ProviderTokens, 7dShare
DeepSeek (lab)11.5T14.3%
DeepInfra10.2T12.7%
Wafer8.93T11.1%
Together7.62T9.5%
Relace7.51T9.3%
inference.net4.80T6.0%
Morph4.34T5.4%
DigitalOcean4.12T5.1%
Open Inference2.02T2.5%
CoreWeave1.67T2.1%
Fireworks1.65T2.1%
StreamLake1.62T2.0%
Crusoe1.35T1.7%
Baidu Qianfan1.31T1.6%
Alibaba Cloud Int.1.05T1.3%
32 others10.8T13.4%

All DeepSeek models combined, Nov 2 to Nov 8, 2026.

DeepSeek V4.1 Flash 1017: prices and speed by provider

Output prices run from $0.15 to $2.40 per million tokens across 40 providers, median $0.675.

ProviderShareInput $/MOutput $/MCache read $/MTokens/sLatencyUptime 3dContext
DeepSeek20.3%$0.15$0.60$0.00301141.30s98.9%1.05M
DeepInfra13.4%$0.078$0.235$0.0024472.84s100.0%1.05M
Together9.9%$0.30$1.20$0.00601880.56s100.0%1.05M
Wafer9.1%$0.038$0.50$0.011350.42s100.0%1.05M
Relace9.0%$0.0060$0.50$0.0050790.82s100.0%1.05M
inference.net7.3%$0.08$0.55$0.00801130.26s100.0%1.04M
Morph5.5%$0.022$0.40$0.0040840.98s99.7%1.05M
DigitalOcean4.4%$0.18$0.72$0.00601510.88s100.0%1.05M
Open Inference2.2%$0.024$0.70$0.00501160.89s100.0%1.05M
Fireworks1.8%$0.21$0.84$0.0042682.67s96.8%1.05M
Particle AI1.3%$0.025$0.25$0.0015690.78s100.0%1.05M
Ionstream1.2%$0.083$0.247$0.0052971.31s-1.05M
Alibaba Cloud Int.1.2%$0.075$0.30$0.0075932.18s100.0%1M
Crusoe1.1%$0.29$1.20$0.00701750.86s100.0%1.05M
StepFun1.1%$0.135$0.54$0.00271311.04s100.0%1M
StreamLake1.0%$0.111$0.444$0.0022621.32s100.0%1.02M
Cloudflare1.0%$0.10$0.60$0.0030881.98s100.0%1.05M
Cohere1.0%$0.195$0.66$0.00602170.36s100.0%1.05M
Parasail0.9%$0.30$1.20$0.00601260.76s100.0%1.05M
Modal0.9%$0.25$1.00$0.00501250.83s100.0%1.05M
Prime Intellect0.8%$0.044$0.70$0.0024651.17s100.0%1.05M
NovitaAI0.8%$0.165$0.66$0.0033711.65s100.0%1.05M
Baseten0.7%$0.30$1.20$0.00701450.48s99.6%1.05M
SiliconFlow0.6%$0.30$0.55$0.0060810.82s100.0%1.05M
ModelRun [by Modular]0.6%$0.17$0.67$0.00601170.94s100.0%1.05M
Sail Research0.5%$0.101$0.15$0.0067--27.5%1.05M
Decart0.4%$0.123$0.492$0.00251011.69s99.8%1.05M
Baidu Qianfan0.4%$0.28$1.13$0.0060762.89s99.7%1.05M
CoreWeave0.3%$0.30$1.20$0.00602110.47s97.9%1.05M
Mistral0.3%$0.30$1.20$0.0070534.69s99.8%1.05M
DekaLLM0.2%$0.18$0.68$0.0060761.85s99.9%1.05M
AtlasCloud0.2%$0.195$0.78$0.0039741.97s100.0%1.05M
Makora0.1%$0.15$1.20$0.00601450.71s100.0%1.05M
GMICloud0.1%$0.18$0.72$0.00361191.67s100.0%1.05M
Reka AI<0.1%$0.16$1.10$0.0080971.24s99.8%1.05M
Nebius Token Factory<0.1%$0.30$1.20$0.0060961.23s100.0%1.05M
Venice<0.1%$0.30$1.20$0.00751021.60s100.0%1M
Phala<0.1%$0.20$1.20$0.0060731.59s100.0%1.05M
PrimaLabs<0.1%$0.24$0.96$0.00481371.99s98.7%1.05M
NEAR AI<0.1%$0.30$1.20$0.0060420.55s100.0%1.05M

List prices as of November 8, 2026, cheapest endpoint per provider. Tokens/s and latency are medians over 30 minutes.

DeepSeek V4.1 Flash: prices and speed by provider

Output prices run from $0.30 to $1.80 per million tokens across 30 providers, median $0.84.

ProviderShareInput $/MOutput $/MCache read $/MTokens/sLatencyUptime 3dContext
Wafer15.8%$0.051$0.567$0.0181140.58s100.0%1.05M
Together11.8%$0.30$1.20$0.00602110.26s100.0%1.05M
DeepInfra11.4%$0.112$0.336$0.0034--41.5%1.05M
Relace8.6%$0.0060$0.50$0.0050740.94s99.5%1.05M
Morph6.9%$0.022$0.40$0.0040580.92s99.8%1.05M
CoreWeave6.8%$0.13$0.65$0.0020900.75s100.0%1.05M
DigitalOcean6.7%$0.225$0.90$0.00601160.89s100.0%1.05M
Open Inference4.4%$0.024$0.70$0.00501120.87s100.0%1.05M
inference.net4.0%$0.103$0.60$0.0080810.41s100.0%1.04M
Fireworks3.5%$0.21$0.84$0.00421060.94s99.8%1.05M
Baseten3.3%$0.30$1.20$0.00701510.94s99.4%1.05M
Baidu Qianfan2.8%$0.30$1.20$0.0060541.21s100.0%1.05M
Crusoe2.8%$0.29$1.20$0.00701190.63s100.0%1.05M
StreamLake2.7%$0.147$0.588$0.0029860.91s100.0%1.02M
Alibaba Cloud Int.2.1%$0.075$0.30$0.00751070.49s100.0%1M
Prime Intellect1.9%$0.039$1.20$0.0012980.69s99.6%1.05M
Parasail1.2%$0.30$1.20$0.00601770.32s100.0%1.05M
Venice0.9%$0.30$1.20$0.00751301.10s99.3%1M
Cohere0.6%$0.195$0.66$0.00301220.50s99.9%1.05M
NovitaAI0.5%$0.195$0.78$0.0039950.95s100.0%1.05M
Makora0.4%$0.26$1.17$0.0058880.53s100.0%1.05M
AtlasCloud0.2%$0.135$0.54$0.013970.63s100.0%1.05M
Reka AI0.2%$0.12$0.96$0.0080660.62s99.7%1.05M
GMICloud0.1%$0.18$0.72$0.0036801.44s100.0%1.05M
Nebius Token Factory0.1%$0.30$1.20$0.01710.77s99.9%1.05M
SiliconFlow<0.1%$0.30$1.20$0.0060521.38s99.9%1.05M
Modal<0.1%$0.30$1.20$0.031320.46s100.0%1.05M
Phala<0.1%$0.21$0.84$0.0042910.96s99.9%1.05M
Mancer<0.1%$0.125$0.50$0.05431.00s99.9%1.05M
Google Vertex-$0.30$1.20$0.03521.17s98.1%1.05M

List prices as of November 8, 2026, cheapest endpoint per provider. Tokens/s and latency are medians over 30 minutes.

All DeepSeek models

ModelTokens, 7dProvidersOutput $/MContextReleased
DeepSeek V4.1 Flash 101756.0T40$0.15 to $2.401.05M2026-10-17
DeepSeek V4.1 Flash17.6T30$0.30 to $1.801.05M2026-09-10
DeepSeek V4 Flash 07313.29T26$0.14 to $1.321.05M2026-07-31
DeepSeek V4 Flash 04232.28T14$0.165 to $0.561.05M2026-04-24
DeepSeek V4 Pro 0813774B21$1.74 to $8.001.05M2026-08-12
DeepSeek V4 Pro 0423569B13$1.50 to $3.831.05M2026-04-24
DeepSeek V3.2286B7$0.31 to $1.00164K2025-12-01
DeepSeek V4 Flash Vision Exp69B2$0.647 to $1.321.05M2026-08-21
DeepSeek V3.151B4$0.95 to $1.50164K2025-08-21
DeepSeek V3 032421B2$1.00 to $1.14164K2025-03-24
DeepSeek V319B2$0.712 to $1.03164K2024-12-26
DeepSeek V3.2 Exp12B1$0.41164K2025-09-29
R1 052811B3$2.15 to $2.29164K2025-05-28
DeepSeek V3.1 Terminus5B1$1.00164K2025-09-22

Finance your GPUs

Serving DeepSeek on your own GPUs? Tell us the cluster, the models you serve and the contracts behind them, and we will share it with partner funders that fit.

Prefer email? hello@amcompute.com