DeepSeek has officially launched the DeepSeek-V4-Pro version. Along with this latest release, DeepSeek announced an update to the API pricing structure introducing Peak Hours and Off-Peak Hours schemes.
This new pricing policy takes effect on August 17, 2026 at 00:00 Beijing Time (or August 16, 2026 at 16:00 UTC). Under the new rules, API usage during off-peak hours will receive a 50% discount compared to peak rates.
Model Specifications and Key Features
DeepSeek-V4 comes in two main variants: DeepSeek-V4-Flash for speed efficiency and DeepSeek-V4-Pro for high reasoning performance. Both models support Thinking Mode and Non-Thinking Mode (where thinking mode is enabled by default).
Model Detail Summary
| Parameter / Feature | DeepSeek-V4-Flash | DeepSeek-V4-Pro |
|---|---|---|
| Model Version | DeepSeek-V4-Flash-0731 |
DeepSeek-V4-Pro-0813 |
| Base URL (OpenAI) | https://api.deepseek.com |
https://api.deepseek.com |
| Base URL (Anthropic) | https://api.deepseek.com/anthropic |
https://api.deepseek.com/anthropic |
| Context Length | 1,000,000 tokens (1M) | 1,000,000 tokens (1M) |
| Max Output Tokens | Up to 384,000 tokens (384K) | Up to 384,000 tokens (384K) |
| Concurrency Limit | 2,500 | 500 |
| Supported Features | JSON Output, Tool Calls, Responses API, Anthropic API, Chat Prefix Completion (Beta), FIM Completion (Beta)* | JSON Output, Tool Calls, Responses API, Anthropic API, Chat Prefix Completion (Beta), FIM Completion (Beta)* |
*\FIM Completion (Beta) feature is only available in Non-Thinking Mode.
API Pricing Structure Changes (Per 1M Tokens)
Billing is based on the accumulated number of input and output tokens. The Context Caching feature continues to provide significant savings for input tokens successfully stored in cache (Cache Hit).
1. Current Standard Pricing
| Model Variant | 1M Input (Cache Hit) | 1M Input (Cache Miss) | 1M Output Tokens |
|---|---|---|---|
| DeepSeek-V4-Flash | $0.0028 | $0.14 | $0.28 |
| DeepSeek-V4-Pro | $0.003625 | $0.435 | $0.87 |
2. New Pricing: Peak & Off-Peak Scheme
Effective August 16, 2026 at 16:00 UTC (August 17, 2026 at 00:00 WIB/Beijing), billing will be adjusted based on the time of API calls.
Peak Hours Schedule:
- 01:00 โ 04:00 UTC (08:00 โ 11:00 WIB)
- 06:00 โ 10:00 UTC (13:00 โ 17:00 WIB)
All hours outside this range are categorized as Off-Peak Hours with 50% cheaper rates.
Pricing Table Based on Usage Hours
| Model | Hour Status | Input (Cache Hit) / 1M | Input (Cache Miss) / 1M | Output / 1M |
|---|---|---|---|---|
| deepseek-v4-flash | OFF-PEAK | $0.007 | $0.22 | $0.66 |
| PEAK | $0.014 | $0.44 | $1.32 | |
| deepseek-v4-pro | OFF-PEAK | $0.022 | $0.66 | $1.98 |
| PEAK | $0.044 | $1.32 | $3.96 |
Balance Deduction Rules & Saving Tips
- Automatic Deduction System: Costs are calculated using the formula Cost = Number of Tokens x Price. This cost is deducted directly from the deposit balance (topped-up balance) or bonus balance (granted balance). If both are available, the system prioritizes deduction from the granted balance.
- Call Time Optimization (Batch Jobs): For large workloads such as mass document processing, data extraction, or indexing, schedule execution during off-peak hours to cut operational costs by up to half.
- Leverage Context Caching: Maximize the use of repetitive prompts (long system prompts or reference documents) to trigger Cache Hit at a much more economical rate.
- Check Prices Periodically: DeepSeek reserves the right to adjust product prices at any time. Developers are advised to top up their balance based on actual needs and regularly monitor the official pricing documentation page.
Related Articles
Getting to Know FFmpeg: The Magic Tool for Video and Audio Processing
Written by
Wilan
A regular contributor to Bali Island Tekno who actively shares knowledge about technology, programming, and the world of software engineering.