DeepSeek · DeepSeek V4
DeepSeek V4 Pro 0813
deepseek/deepseek-v4-pro-0813
The hosted V4 Pro snapshot currently served by DeepSeek's official API, with million-token context, 384K output and peak/off-peak billing.
Hosted weightsReasoningTool callingStructured outputTemperature
Model specifications
- Context window
- 1M tokens
- Output limit
- 384K tokens
- Release date
- 2026-08-12
- Input modalities
- text
- Output modalities
- text
- Last updated
- 2026-08-13
Provider pricing
USD per 1M tokens. Limits may differ from the underlying model.
Checked 2026-08-17
DeepSeek API (off-peak)
deepseek-v4-pro
- Input
- $0.66
- Output
- $1.98
- Context
- 1M
DeepSeek API (peak)
deepseek-v4-pro
- Input
- $1.32
- Output
- $3.96
- Context
- 1M
Promotions, plan quotas, cache pricing and regional taxes are not included. Recheck the provider before purchase.
Sources
DeepSeek models and pricingConfirms the V4-Pro-0813 route, 1M context, 384K output, feature support, concurrency and peak/off-peak prices. Offering input prices use cache-miss rates; cache-hit rates are $0.022/M off-peak and $0.044/M peak.Models.dev canonical model recordProvides a normalized record for the hosted 0813 snapshot and its capabilities and limits.