请确保 OpenAI 库版本为 1.0.0 或更高；否则，请参阅旧文档 OpenAI 适配器（旧版）。 许多人从 OpenAI 开始，但想探索其他模型。LangChain 与许多模型提供程序的集成使这变得容易。虽然 LangChain 有自己的消息和模型 API，但我们也通过公开一个适配器将 LangChain 模型适配到 OpenAI API，使探索其他模型变得尽可能容易。 目前这仅处理输出，不返回其他信息（token 计数、停止原因等）。 
import openai
from langchain_community.adapters import openai as lc_openai

​
chat.completions.create

messages = [{"role": "user", "content": "hi"}]
原始 OpenAI 调用 
result = openai.chat.completions.create(
    messages=messages, model="gpt-3.5-turbo", temperature=0
)
result.choices[0].message.model_dump()

{'content': 'Hello! How can I assist you today?',
 'role': 'assistant',
 'function_call': None,
 'tool_calls': None}
LangChain OpenAI 包装器调用 
lc_result = lc_openai.chat.completions.create(
    messages=messages, model="gpt-3.5-turbo", temperature=0
)

lc_result.choices[0].message  # Attribute access

{'role': 'assistant', 'content': 'Hello! How can I help you today?'}

lc_result["choices"][0]["message"]  # Also compatible with index access

{'role': 'assistant', 'content': 'Hello! How can I help you today?'}
交换模型提供程序 
lc_result = lc_openai.chat.completions.create(
    messages=messages, model="claude-2", temperature=0, provider="ChatAnthropic"
)
lc_result.choices[0].message

{'role': 'assistant', 'content': 'Hello! How can I assist you today?'}

​
chat.completions.stream

原始 OpenAI 调用 
for c in openai.chat.completions.create(
    messages=messages, model="gpt-3.5-turbo", temperature=0, stream=True
):
    print(c.choices[0].delta.model_dump())

{'content': '', 'function_call': None, 'role': 'assistant', 'tool_calls': None}
{'content': 'Hello', 'function_call': None, 'role': None, 'tool_calls': None}
{'content': '!', 'function_call': None, 'role': None, 'tool_calls': None}
{'content': ' How', 'function_call': None, 'role': None, 'tool_calls': None}
{'content': ' can', 'function_call': None, 'role': None, 'tool_calls': None}
{'content': ' I', 'function_call': None, 'role': None, 'tool_calls': None}
{'content': ' assist', 'function_call': None, 'role': None, 'tool_calls': None}
{'content': ' you', 'function_call': None, 'role': None, 'tool_calls': None}
{'content': ' today', 'function_call': None, 'role': None, 'tool_calls': None}
{'content': '?', 'function_call': None, 'role': None, 'tool_calls': None}
{'content': None, 'function_call': None, 'role': None, 'tool_calls': None}
LangChain OpenAI 包装器调用 
for c in lc_openai.chat.completions.create(
    messages=messages, model="gpt-3.5-turbo", temperature=0, stream=True
):
    print(c.choices[0].delta)

{'role': 'assistant', 'content': ''}
{'content': 'Hello'}
{'content': '!'}
{'content': ' How'}
{'content': ' can'}
{'content': ' I'}
{'content': ' assist'}
{'content': ' you'}
{'content': ' today'}
{'content': '?'}
{}
交换模型提供程序 
for c in lc_openai.chat.completions.create(
    messages=messages,
    model="claude-2",
    temperature=0,
    stream=True,
    provider="ChatAnthropic",
):
    print(c["choices"][0]["delta"])

{'role': 'assistant', 'content': ''}
{'content': 'Hello'}
{'content': '!'}
{'content': ' How'}
{'content': ' can'}
{'content': ' I'}
{'content': ' assist'}
{'content': ' you'}
{'content': ' today'}
{'content': '?'}
{}
Edit the source of this page on GitHub.
Connect these docs programmatically to Claude, VSCode, and more via MCP for real-time answers.