DeepSeek V4.1 Flash
deepseek-flash · V4.1-Flash-1017| Token type | Off-peak | Peak |
|---|---|---|
| Input, cache miss | $0.15 | $0.30 |
| Input, cache hit | $0.003 | $0.006 |
| Output | $0.60 | $1.20 |
AI cost tool / DeepSeek API
DeepSeek prices the same tokens differently depending on when a request runs. This page turns that schedule into a live rate board: what is active now, what comes next, and the cost per million tokens.
Rates / per 1M tokens
Off-peak is exactly half of peak. Both bands are shown; the current one is highlighted.
| Token type | Off-peak | Peak |
|---|---|---|
| Input, cache miss | $0.15 | $0.30 |
| Input, cache hit | $0.003 | $0.006 |
| Output | $0.60 | $1.20 |
| Token type | Off-peak | Peak |
|---|---|---|
| Input, cache miss | $0.66 | $1.32 |
| Input, cache hit | $0.022 | $0.044 |
| Output | $1.98 | $3.96 |
Prices in USD. Tool fees such as web search are separate and do not change with the price window.
Schedule / UTC
Peak runs Monday to Friday, 01:00 to 04:00 and 06:00 to 10:00 UTC. Everything else is off-peak, including all day Saturday and Sunday.
Hours are UTC. On a phone, scroll the week sideways. Weekends are off-peak all day.
Monday to Friday, 01:00-04:00 and 06:00-10:00 UTC.
Every other hour, plus all day Saturday and Sunday.
Chinese public holidays are off-peak. This clock uses the weekly schedule and does not subtract holiday dates.
Control the cost
We build AI workflows with human review, guardrails, and a cost model you can see. If your AI work needs a real operating layer, start here.
Pricing is read from DeepSeek's published model table. Image inputs follow its Vision billing rules, and the concurrency values come from its rate-limit documentation. DeepSeek can change prices or limits at any time, so confirm the final number in your own account before committing to a budget.
Independent tool. Not affiliated with DeepSeek.