DeepSeek’s API is priced per million tokens, and the price depends on the time of day. Its documentation (read 29 Sep 2026) lists two models, deepseek-flash (DeepSeek-V4.1-Flash) and deepseek-v4-pro, each with a 1M-token context and up to 384K tokens of output. Flash costs $0.15 per million input tokens at off-peak and $0.30 at peak; V4 Pro costs $0.66 and $1.32. Off-peak is exactly half of peak. The V3 and R1 prices this page used to show are no longer on the pricing page.
Price per million tokens
| Model | Input, cache hit | Input, cache miss | Output |
|---|---|---|---|
| deepseek-flash, off-peak | $0.003 | $0.15 | $0.60 |
| deepseek-flash, peak | $0.006 | $0.30 | $1.20 |
| deepseek-v4-pro, off-peak | $0.022 | $0.66 | $1.98 |
| deepseek-v4-pro, peak | $0.044 | $1.32 | $3.96 |
Peak hours are 01:00 to 04:00 and 06:00 to 10:00 UTC, Monday to Friday, excluding Chinese public holidays. Everything else, including weekends, is off-peak. The old model names deepseek-v4-flash and deepseek-v4-flash-vision-exp are listed as retired but still accepted.
Worked cost example
Our arithmetic: 100 requests, each with 2,000 input tokens (none cached) and 500 output tokens, so 200,000 input and 50,000 output tokens.
| Model | Peak | Off-peak |
|---|---|---|
| deepseek-flash | $0.12 | $0.06 |
| deepseek-v4-pro | $0.46 | $0.23 |
Input at cache-hit rates is a small fraction of these figures, so repeated long prompts cost far less than the table implies.
Who it suits, who should skip
- Suits developers who call an API at volume and can schedule batch work for off-peak hours.
- Skip if you want a consumer subscription with a plan table: the documentation covers the API only. Compare ChatGPT, Claude and Gemini.
- Check before adopting where your data goes and whether your organisation permits a China-based provider; we have not reviewed DeepSeek’s terms or data policy.
- For open-weight image models, see Flux.1; for another Europe-based assistant, Mistral Le Chat.
What this page cannot tell you
Answer quality, speed and reliability are not on a pricing page, and we have not assessed them. We did not verify whether a free web chat is still offered, or its licence terms, so the earlier statements on both were removed.
Method
Researched from DeepSeek’s API documentation, read 29 Sep 2026. USD per million tokens as displayed; the cost example is our arithmetic. Nothing here comes from using the product.