DeepSeek · API model
DeepSeek V4 Pro
API id deepseek-v4-pro prices checked
- Context
- 1.05Mtokens
- Max output
- 384Ktokens
- Knowledge cutoff
- Not disclosed
- Typical request
- $0.002
Pricing
US dollars per 1 million tokens. A typical request (1,000 input + 500 output tokens) costs $0.002 at standard rates.
| Rate | Input | Cached input | Output |
|---|---|---|---|
| Standard | $0.66 | $0.022 | $1.98 |
Sources list different standard prices. DeepSeek bills by time of day. The headline on this page is DeepSeek's own off-peak rate ($0.66 / $1.98). During peak hours (01:00-04:00 and 06:00-10:00 UTC, Monday to Friday) the price doubles to $1.32 / $3.96. OpenRouter, a third-party router, lists a flat $0.58 / $1.74.
| Source | Input | Output | Note |
|---|---|---|---|
| vendorDeepSeek API docs | $0.66 | $1.98 | Off-peak, used on this page |
| vendorDeepSeek API docs | $1.32 | $3.96 | Peak: 01:00-04:00 and 06:00-10:00 UTC, Mon-Fri |
| independentOpenRouter | $0.58 | $1.74 | Flat rate via a third-party router |
Capabilities and limits
| Context window | 1.05M tokens |
|---|---|
| Max output | 384K tokens |
| Input | Text |
| Output | Text |
| Reasoning | Adjustable effort |
| Tool use | Yes |
| Structured output (JSON) | Yes |
| Knowledge cutoff | Not disclosed by the vendor |
Change history
Launches, price changes and retirements for DeepSeek V4 Pro, newest first, each linked to the vendor's notice.
DeepSeek V4.1 Flash released with new peak and off-peak prices
deepseek-flash costs $0.15 / $0.60 off-peak and $0.30 / $1.20 at peak. V4 Flash and V4-Flash-Vision-Exp are retired; their names route to V4.1 Flash. A planned move of V4 Pro traffic to Flash from September 14 was later withdrawn, and V4 Pro stays at unchanged prices.
DeepSeek V4 Pro generally available
The model version behind deepseek-v4-pro is DeepSeek-V4-Pro-0813.
vendorDeepSeek news
Benchmarks and speed
Not collected yet. Scores will come only from independent leaderboards, each shown with its source and test date.
Planned: GPQA Diamond, SWE-bench Verified, AIME 2025, LMArena Elo, output speed and time to first token.
Where these numbers come from
Every value is an observation with a source and a date. independent is a third-party catalog, vendor is the model maker's own page, derived is our calculation.
| Field | Value | Source | Observed |
|---|---|---|---|
| Off-peak price, input / cached / output | $0.66 / $0.022 / $1.98 | vendorDeepSeek API docs | Sep 16, 2026 |
| Peak price, input / cached / output | $1.32 / $0.044 / $3.96 | vendorDeepSeek API docs | Sep 16, 2026 |
| Router price, input / output | $0.58 / $1.74 | independentOpenRouter | Sep 16, 2026 |
| Context window | 1,048,576 | independentOpenRouter | Sep 15, 2026 |
| Max output | 384K | vendorDeepSeek API docs | Sep 16, 2026 |
| Reasoning | Hybrid: non-think, think-high, think-max modes | vendorDeepSeek docs | Sep 15, 2026 |
| Release date | 2026-04-24 (preview) | vendorDeepSeek API docs | Sep 15, 2026 |
| Typical request | $0.0017 | derivedCalculated: 1,000 in + 500 out at off-peak price | Sep 16, 2026 |