modelslab-chat-generation

ModelsLab 聊天/大模型生成

通过单个兼容 OpenAI 的端点，向 60 多种大模型发送聊天补全请求。

何时使用此技能

• 与 AI 模型（DeepSeek、Llama、Gemini、Qwen、Mistral）聊天
• 构建对话式 AI 应用
• 使用系统提示词生成文本补全
• 使用大模型进行函数/工具调用
• 流式传输响应，实现实时输出
• 获取结构化 JSON 响应

API 端点

聊天补全：POST https://modelslab.com/api/v7/llm/chat/completions

该端点遵循 OpenAI 聊天补全格式，因此可以轻松从 OpenAI 切换，或使用 OpenAI SDK。

快速开始

import requests

def chat(message, api_key, model="meta-llama-3-8B-instruct"):
    """Send a chat completion request.

    Args:
        message: The user message
        api_key: Your ModelsLab API key
        model: LLM model ID (use modelslab models search --feature llmaster to find models)
    """
    response = requests.post(
        "https://modelslab.com/api/v7/llm/chat/completions",
        json={
            "key": api_key,
            "model_id": model,
            "messages": [
                {"role": "user", "content": message}
            ],
            "max_tokens": 200,
            "temperature": 0.7
        }
    )

    data = response.json()

    if "choices" in data:
        return data["choices"][0]["message"]["content"]
    else:
        raise Exception(f"Error: {data.get('message', 'Unknown error')}")

# Usage
reply = chat(
    "Explain quantum computing in simple terms.",
    "your_api_key"
)
print(reply)

使用系统提示词

def chat_with_system(system_prompt, message, api_key, model="meta-llama-3-8B-instruct"):
    """Chat with a system prompt for role/behavior control."""
    response = requests.post(
        "https://modelslab.com/api/v7/llm/chat/completions",
        json={
            "key": api_key,
            "model_id": model,
            "messages": [
                {"role": "system", "content": system_prompt},
                {"role": "user", "content": message}
            ],
            "max_tokens": 500,
            "temperature": 0.7
        }
    )

    data = response.json()
    if "choices" in data:
        return data["choices"][0]["message"]["content"]

# Usage
reply = chat_with_system(
    "You are a Python expert. Give concise code examples.",
    "How do I read a CSV file?",
    "your_api_key"
)

多轮对话

def conversation(messages, api_key, model="meta-llama-3-8B-instruct"):
    """Send a multi-turn conversation."""
    response = requests.post(
        "https://modelslab.com/api/v7/llm/chat/completions",
        json={
            "key": api_key,
            "model_id": model,
            "messages": messages,
            "max_tokens": 500,
            "temperature": 0.7
        }
    )
    return response.json()

# Multi-turn example
messages = [
    {"role": "system", "content": "You are a helpful assistant."},
    {"role": "user", "content": "What is Python?"},
    {"role": "assistant", "content": "Python is a high-level programming language..."},
    {"role": "user", "content": "Show me a hello world example."}
]

result = conversation(messages, "your_api_key")
print(result["choices"][0]["message"]["content"])

兼容 OpenAI SDK

更改基础 URL 和 API 密钥后，即可使用官方 OpenAI Python SDK：

from openai import OpenAI

client = OpenAI(
    base_url="https://modelslab.com/api/v7/llm",
    api_key="your_modelslab_api_key"
)

response = client.chat.completions.create(
    model="meta-llama-3-8B-instruct",
    messages=[
        {"role": "system", "content": "You are a helpful assistant."},
        {"role": "user", "content": "Hello!"}
    ],
    max_tokens=100,
    temperature=0.7
)

print(response.choices[0].message.content)

cURL 示例

curl -X POST "https://modelslab.com/api/v7/llm/chat/completions" \
  -H "Content-Type: application/json" \
  -d '{
    "key": "your_api_key",
    "model_id": "meta-llama-3-8B-instruct",
    "messages": [
      {"role": "user", "content": "Say hello in one sentence."}
    ],
    "max_tokens": 50,
    "temperature": 0.7
  }'

发现大模型

# Search all chat/LLM models
modelslab models search --feature llmaster

# Search by provider
modelslab models search --search "deepseek"
modelslab models search --search "llama"
modelslab models search --search "gemini"

# Get model details
modelslab models detail --id meta-llama-3-8B-instruct

常用大模型 ID

Meta Llama

• meta-llama-3-8B-instruct — Llama 3 8B（快速、高效）
• meta-llama-Llama-3.3-70B-Instruct-Turbo — Llama 3.3 70B（功能强大）
• meta-llama-Meta-Llama-3.1-405B-Instruct-Turbo — Llama 3.1 405B（规模最大）

DeepSeek

• deepseek-ai-DeepSeek-R1-Distill-Llama-70B — DeepSeek R1（推理）
• deepseek-ai-DeepSeek-V3 — DeepSeek V3（通用）

Google

• gemini-2.0-flash-001 — Gemini 2.0 Flash（快速）
• gemini-2.5-pro — Gemini 2.5 Pro（能力强）

Qwen

• Qwen-Qwen2.5-72B-Instruct-Turbo — Qwen 2.5 72B
• Qwen-Qwen2.5-Coder-32B-Instruct — Qwen 2.5 Coder

Mistral

• mistralai-Mixtral-8x7B-Instruct-v0.1 — Mixtral 8x7B
• mistralai-Mistral-Small-24B-Instruct-2501 — Mistral Small

关键参数

参数	类型	必填	描述
model_id	string	是	大模型标识符（别名：model）
messages	array	是	包含 role 和 content 的消息对象数组
temperature	float	否	采样温度（0-2，默认值因模型而异）
max_tokens	integer	否	要生成的最大 token 数
top_p	float	否	核采样（0-1）
top_k	integer	否	Top-k 采样
frequency_penalty	float	否	频率惩罚（-2 到 2）
presence_penalty	float	否	存在惩罚（-2 到 2）
streaming	boolean	否	启用流式响应（别名：stream）
n	integer	否	补全数量（1-10）
stop	array	否	停止序列（最多 4 个）
seed	integer	否	用于确保结果可复现的随机种子
tools	array	否	函数/工具定义
tool_choice	string	否	工具选择策略
response_format	object	否	JSON 模式使用 {"type": "json_object"}

响应格式

响应遵循 OpenAI 聊天补全格式：

{
  "id": "gen-...",
  "model": "meta-llama-3-8B-instruct",
  "object": "chat.completion",
  "created": 1771658583,
  "choices": [
    {
      "index": 0,
      "finish_reason": "stop",
      "message": {
        "role": "assistant",
        "content": "Hello! How can I help you today?"
      }
    }
  ],
  "usage": {
    "prompt_tokens": 17,
    "completion_tokens": 10,
    "total_tokens": 27
  }
}

错误处理

def safe_chat(message, api_key, model="meta-llama-3-8B-instruct"):
    """Chat with error handling."""
    try:
        response = requests.post(
            "https://modelslab.com/api/v7/llm/chat/completions",
            json={
                "key": api_key,
                "model_id": model,
                "messages": [{"role": "user", "content": message}],
                "max_tokens": 200
            }
        )
        data = response.json()

        if "choices" in data:
            return data["choices"][0]["message"]["content"]
        elif data.get("status") == "error":
            raise Exception(data.get("message", "Unknown error"))
        else:
            raise Exception(f"Unexpected response: {data}")

    except requests.exceptions.RequestException as e:
        raise Exception(f"Network error: {e}")

try:
    reply = safe_chat("Hello!", "your_api_key")
    print(reply)
except Exception as e:
    print(f"Chat failed: {e}")

资源

• API 文档：https://docs.modelslab.com/uncensored-chat-api/overview
• 模型浏览器：https://modelslab.com/models
• 获取 API 密钥：https://modelslab.com/dashboard/api-keys
• 命令行工具：https://github.com/ModelsLab/modelslab-cli

相关技能

• modelslab-model-discovery - 查找和筛选模型
• modelslab-account-management - 管理 API 密钥和账单
• modelslab-image-generation - 根据文本生成图像