Model Explorer
Technical Specifications
Detailed technical parameters and capabilities
- Context Window
- 1,000,000 tokens
- Max Output
- 384,000 tokens
- Training Cutoff
- 2026-04
- Tokenizer
- DeepSeek BPE
Supported Modes
Chat
Capabilities
Chat
Multi-turn conversational completions
Streaming
Tokens are delivered incrementally as they are generated
Tool Use
Can call functions/tools defined in the request
Thinking
Produces an extended internal reasoning trace before the final answer
Prompt Caching
Reuses previously processed prompt prefixes at a lower rate
Quick Integration
Get started with a single API call
quickstart.py
import openai
client = openai.OpenAI(
base_url="https://api.example.com/v1",
api_key="sk-ms-...",
)
response = client.chat.completions.create(
model="deepseek-v4-flash-0731",
messages=[{"role": "user", "content": "Hello!"}],
)
print(response.choices[0].message.content)
Pay-as-you-go
−57%
Cost
No commitments, pay only for what you use
Full Pricing Breakdown
Every billed dimension for deepseek-v4-flash-0731
Token Pricing
- Cache Read$0.00595 per 1M tokens
- Input$0.187 per 1M tokens
- Output$0.561 per 1M tokens
Indicative pricing — final cost is confirmed at request time.
Time-based Pricing
- Defaultall other times$0.187 / $0.561per 1M tokens
- DeepSeek Peak AMMon–Fri 01:00–04:00 UTC$0.374 / $1.122per 1M tokens
- DeepSeek Peak LateMon–Fri 06:00–10:00 UTC$0.374 / $1.122per 1M tokens
Window times follow the timezone shown; requests crossing a window boundary are billed by the window the request started in.