Provider pricing
DeepSeek API pricing: every model, cheapest provider
23 DeepSeek models are offered by 20 providers we track. The lowest listed rate is deepseek-r1-distill-llama-8b at $0.025 input and $0.025 output per 1M tokens from Nscale.
Standard list prices in USD. Last price change Nov 1, 2026. How we collect and rank prices.
Open a model to compare every provider
DeepSeek models and lowest prices
| Model | Providers | Cheapest provider | Input / 1M | Output / 1M | Max context |
|---|---|---|---|---|---|
| DeepSeek V4 Flash 0731 | 10 | DeepInfra | $0.08 | $0.18 | 1M |
| DeepSeek V4 Pro 0813 | 8 | DeepInfra | $1.30 | $2.60 | 1M |
| DeepSeek-R1 | 8 | SiliconFlow | $0.50 | $2.18 | 164K |
| DeepSeek-V3 | 8 | SiliconFlow | $0.25 | $1.00 | 164K |
| DeepSeek-V3.2 | 8 | DeepInfra | $0.26 | $0.38 | 164K |
| DeepSeek-V4.1-Flash | 8 | DeepInfra | $0.14 | $0.42 | 1M |
| DeepSeek V4.1 Flash 1017 | 7 | DeepInfra | $0.14 | $0.42 | 1M |
| DeepSeek-V3.1 | 7 | DeepInfra | $0.25 | $0.95 | 164K |
| DeepSeek-V4-Flash | 7 | DeepInfra | $0.09 | $0.18 | 1M |
| DeepSeek V4 Pro | 6 | DeepInfra | $1.30 | $2.60 | 1M |
| DeepSeek-R1-0528 | 6 | Lambda | $0.20 | $0.60 | 164K |
| deepseek-r1-distill-llama-70b | 6 | DeepInfra | $0.20 | $0.60 | 131K |
| Deepseek R1 Distill Qwen 32B | 5 | Nscale | $0.15 | $0.15 | 131K |
| deepseek-v3-0324 | 5 | Lambda | $0.20 | $0.60 | 164K |
| DeepSeek V4 Flash Vision Exp | 4 | DeepSeek | $0.30 | $1.20 | 1M |
| DeepSeek-V3.1-Terminus | 4 | DeepInfra | $0.27 | $1.00 | 164K |
| deepseek-r1-distill-qwen-14b | 3 | Nscale | $0.07 | $0.07 | 131K |
| DeepSeek R1 (Turbo) | 2 | Novita AI | $0.70 | $2.50 | 64K |
| DeepSeek V4.1 Flash 1017 | 2 | DeepSeek | $0.30 | $1.20 | 1M |
| deepseek-r1-distill-llama-8b | 2 | Nscale | $0.025 | $0.025 | 131K |
| deepseek-r1-distill-qwen-1.5b | 2 | Nscale | $0.09 | $0.09 | 131K |
| deepseek-r1-distill-qwen-7b | 2 | Fireworks AI | $0.20 | $0.20 | 131K |
| DeepSeek-V3.2-Exp | 2 | Novita AI | $0.27 | $0.41 | 164K |
Only models listed by at least three providers are shown. Prices are per 1M tokens at the cheapest provider for each model.
Compare more
Other model developers
Alibaba Qwen, Anthropic Claude, Google Gemini and Gemma, Meta Llama, MiniMax, Mistral AI, Moonshot AI Kimi, NVIDIA Nemotron, OpenAI, xAI Grok, Z.ai GLM