DeepSeek V3
Create a chat completion for given messages with streaming support using DeepSeek’s V3 model, optimized for reasoning, mathematics, coding, and complex analytical tasks
Documentation Index
Fetch the complete documentation index at: https://docs.shaktistudio.ai/llms.txt
Use this file to discover all available pages before exploring further.
Authorizations
JWT token for authentication - Use your API token as the Bearer token
Headers
Model UUID for the request (DeepSeek V3 model identifier)
Body
Array of messages in the conversation
Model identifier for DeepSeek V3
deepseek-v3 Whether to stream the response
Sampling temperature (0.0 to 2.0). Lower values for analytical tasks, higher for creative work
0 <= x <= 2Maximum number of tokens to generate
1 <= x <= 65536Nucleus sampling parameter for controlling response diversity
0 <= x <= 1Top-k sampling parameter for vocabulary selection
1 <= x <= 100Sequences where the API will stop generating further tokens
4Penalty for frequent tokens to reduce repetition
-2 <= x <= 2Penalty for new tokens to encourage topic diversity
-2 <= x <= 2Penalty for repeating tokens (DeepSeek-specific parameter)
0.1 <= x <= 2Whether to use sampling for generation
Random seed for reproducible outputs
Enable enhanced reasoning capabilities for complex problems
Response
Successful chat completion
- Option 1
- Option 2
Response for non-streaming chat completion
Unique identifier for the completion
"chatcmpl-deepseek-abc123"
Object type
chat.completion "chat.completion"
Unix timestamp of when the completion was created
1699014493
The model used for completion
"deepseek-v3"
System fingerprint for the model version
"deepseek-v3-20241201"

