What are the official DeepSeek API rates and models?

DeepSeek API pricing starts at $0.15 per million uncached input tokens and $0.60 per million output tokens for the official Flash API during off-peak hours, as detailed in the guia de precios de deepseek para desarrolladores. Because other providers set their own rates, the total cost depends heavily on where you access the service and which exact configuration you choose.

The current model behind deepseek-flash is DeepSeek-V4.1-Flash-1017, though older names are still accepted and routed to this model at Flash prices. Developers can explore additional viral repositories like open-yoki to deploy autonomous AI live streamers, or review static references using ruic-poseatlas for design and prompting workflows.

How do peak hours and cache hits impact pricing?

Peak pricing applies from 01:00 to 04:00 UTC and 06:00 to 10:00 UTC, Monday through Friday, excluding Chinese public holidays. Every other hour, including weekends and holidays, qualifies for lower off-peak rates.

When planning monthly budgets, token costs are calculated by multiplying cache-miss inputs, cache-hit inputs, and billable outputs by their respective rates. Reusing a document via prompt caching significantly reduces input costs, dropping the official off-peak rate from $0.15 to just $0.003 per million tokens.

How do provider alternatives and workloads change costs?

Different providers offer distinct rate structures that favor specific workloads. For instance, Morph offers a lower uncached input rate of $0.12 but higher output costs, whereas Moduck provides an FP4 quantization configuration priced at $0.03776 for uncached input and $0.096 for output.

To estimate your monthly expenditure accurately, follow these steps to calculate token expenses and select the best provider:

  1. Gather your expected monthly request volume and average input/output token counts.
  2. Split your input tokens into cache-hit and cache-miss categories based on measured usage.
  3. Apply the provider's specific rates for off-peak and peak hours to each token category.
  4. Add separately billed features such as research agent search and fetch requests to finalize the total budget.