DeepSeek

1. Overview

DeepSeek offers the most affordable domestic large language models currently available, with low prompt/generation costs, making it ideal for translation needs.

This API is compatible with the OpenAI interface format.

Model List:

• deepseek-v3

• deepseek-v3.1

• deepseek-r1

• deepseek-v4-flash

• deepseek-v4-pro

2. Request Description

• Request Method: POST

• Request URL:

  https://gateway.theturbo.ai/v1/chat/completions

To ensure concurrent resource availability, the backend uses multi-account load balancing. To improve cache hit rates in multi-turn conversation mode, include the HTTP request header X-Conversation-Id with a random string in your request. The platform will preferentially route requests to the same backend account. Reference Documentation

---

3. Request Parameters

3.1 Header Parameters

Parameter Name	Type	Required	Description	Example
Content-Type	string	Yes	Set the request header type, must be application/json	application/json
Accept	string	Yes	Set the response type, recommended to use application/json	application/json
Authorization	string	Yes	API_KEY required for authentication, format Bearer $YOUR_API_KEY	Bearer $YOUR_API_KEY

---

3.2 Body Parameters (application/json)

Parameter Name	Type	Required	Description	Example
model	string	Yes	The model ID to use. See Overview for available versions, e.g. deepseek-v3.	deepseek-v3
messages	array	Yes	Chat message list, compatible with OpenAI format. Each object in the array contains role and content.	[{"role": "user","content": "Hello"}]
role	string	No	Message role, possible values: system, user, assistant.	user
content	string	No	The specific content of the message.	Hello, please tell me a joke.
temperature	number	No	Sampling temperature, ranging from 0~2. Higher values make the output more random; lower values make the output more focused and deterministic.	0.7
top_p	number	No	Another way to adjust the sampling distribution, ranging from 0~1. Usually set one of these or temperature.	0.9
n	number	No	Number of replies to generate for each input message.	1
stream	boolean	No	Whether to enable streaming output. When set to true, returns streaming data similar to ChatGPT.	false
stop	string	No	Up to 4 strings can be specified. Once any of these strings appears in the generated content, token generation stops.	"\n"
max_tokens	number	No	Maximum number of tokens that can be generated in a single reply, limited by the model's context length.	1024
presence_penalty	number	No	-2.0 ~ 2.0. Positive values encourage the model to generate more new topics, while negative values reduce the probability of new topics.	0
frequency_penalty	number	No	-2.0 ~ 2.0. Positive values reduce the frequency of the model repeating phrases, while negative values increase the probability of repetition.	0

---

4. Request Examples

HTTP

POST /v1/chat/completions
Content-Type: application/json
Accept: application/json
Authorization: Bearer $YOUR_API_KEY

{
	"model": "deepseek-v3",
	"messages": [
		{
			"role": "user",
			"content": "Hello, can you explain quantum mechanics to me?"
		}
	]
}

Shell

Go

Python

5. Response Example

{
	"id": "chatcmpl-1234567890",
	"object": "chat.completion",
	"created": 1699999999,
	"model": "deepseek-v3",
	"choices": [
		{
			"message": {
				"role": "assistant",
				"content": "Quantum mechanics is a branch of physics that studies the microscopic world..."
			},
			"finish_reason": "stop"
		}
	],
	"usage": {
		"prompt_tokens": 10,
		"completion_tokens": 30,
		"total_tokens": 40
	}
}