DeepSeek API
Usage-based API access to DeepSeek models. Prices differ between peak and off-peak hours, and concurrency limits differ per model.
Platforms: API
Plans and what they cover
Not disclosed means the official page states no figure; not researched means we have not checked the item yet. Neither is the same as "no limit".
Usage-based
Usage-basedUsage-based (no monthly fee)
No monthly equivalent; check the official page for pricing details
Peak hours are 01:00-04:00 and 06:00-10:00 UTC, Monday to Friday; all other times are off-peak. There is no fixed monthly fee.
Models and their allowances
Listed models are those we could confirm. When a model has no individual figure, check the shared plan allowance or separate contract terms as well.
- DeepSeek V4.1 FlashAvailable
The model name is deepseek-flash. Prices differ between off-peak and peak.
- Concurrent requests: 2,500 concurrent
The concurrency cap for each model.
Resets: None / On overage: Not researched
Official source for the per-model allowance
- Concurrent requests: 2,500 concurrent
- DeepSeek V4 ProAvailable
Prices differ between off-peak and peak.
- Concurrent requests: 500 concurrent
A lower concurrency cap than Flash.
Resets: None / On overage: Not researched
Official source for the per-model allowance
- Concurrent requests: 500 concurrent
Usage limit details
- Scope
- Whole plan
- Window
- off-peak
- Resets
- Not researched
- On overage
- Not researched
- Rollover
- Not confirmed
$0.30 at peak, and $0.003 on a cache hit.
Official source for this limitOfficially confirmed · Checked
Official plan page for the latest termsOfficial sources read (1)
- https://api-docs.deepseek.com/quick_start/pricing checked 11/15/2026
Sources for this service
- https://api-docs.deepseek.com/quick_start/pricing checked 11/15/2026