DeepSeek pricing
DeepSeek V4 Pro pricing
DeepSeek V4 Pro bills per million tokens, metered separately for input and output. The current rates are below, along with the parts of the bill a per-token estimate leaves out.
DeepSeek V4 Pro rates in full
All figures are USD per million tokens, taken from DeepSeek’s own pricing documentation. The API model identifier is deepseek-v4-pro.
| Per million tokens | Off-peak (most hours) | Peak (01:00–04:00 & 06:00–10:00 UTC, Mon–Fri) |
|---|---|---|
| Input | $0.66 | $1.32 |
| Cached input read | $0.022 | $0.044 |
| Output | $1.98 | $3.96 |
Context window: 1M tokens, with up to 384K tokens of output per request.
What the rate table does not tell you
DeepSeek is the only provider in this list that varies price by time of day rather than by a fixed rate: peak hours (01:00–04:00 and 06:00–10:00 UTC, Monday to Friday) bill at exactly double the off-peak rate on every line; all other hours are off-peak.
V4 Pro is model version DeepSeek-V4-Pro-0813; DeepSeek confirmed continued API access to it past September 14, 2026 in response to user demand, with billing unchanged.
Cache hits cost about 3% of the cache-miss (standard) input rate in both windows.
A worked monthly example
Take 500,000 calls in a month at 3,000 input tokens and 400 output tokens each. That is 1.5 billion input tokens against 200 million output tokens, so input is 88% of the token volume.
| Line | Tokens | Rate | Cost |
|---|---|---|---|
| Input | 1,500M | $0.66 | $990 |
| Output | 200M | $1.98 | $396 |
| Total | 1,700M | $1,386 |
Output is 12% of the tokens and 29% of the bill. That inversion is why prompt trimming so often fails to move the invoice, and it is the single most common reason a forecast built on DeepSeek V4 Pro comes in under the real number.
How this page stays current
The rate on this page was read off DeepSeek’s published pricing and recorded with the date it was checked, September 16, 2026. The date changes only when someone checks again, not on each deploy. This model is not matched to the live model catalog, so the rate rests on the publisher documentation check alone.
Frequently asked questions
How much does the DeepSeek V4 Pro API cost?
DeepSeek V4 Pro is billed per million tokens, metered separately for input and output. As of September 16, 2026 the rate is $0.66 per million input tokens and $1.98 per million output tokens, with cached input read at $0.022.
Why is my DeepSeek V4 Pro bill higher than the per-token estimate?
Most estimates multiply expected tokens by the input rate and stop there. Output is priced at $1.98 against $0.66 for input, so responses usually dominate the bill even when prompts look larger. Reasoning tokens bill as output. Retries and agent fan-out pay full input cost for calls that produced nothing. And a cache that gets written but never read is pure overhead rather than a saving.
What is the cheapest way to run DeepSeek V4 Pro?
Caching a stable prefix takes repeat input to $0.022, 3.3% of the standard rate, and pays for itself once the cached content is read more often than it is written. The larger lever is usually routing: sending the requests that do not need this tier to a cheaper model, and capping spend per customer before the call goes out.
How do I track DeepSeek V4 Pro spend per customer or per feature?
Provider reports group usage by model, project or API key. Your application supplies the customer and workflow context needed to see cost per account. That means passing the customer, team or agent with each request from your server and recording the resolved cost against it. Spendline does this for supported requests routed through it, and checks configured budgets before forwarding them. Actual response cost can differ from the pre-call estimate.
See cost per customer, not just per token
A rate card tells you what a token costs. It does not tell you which customer, team or agent spent it. Spendline records attributed cost for supported requests routed through it and checks configured budgets before forwarding them.
Book a 30-minute call Explore a sample account