Chat Completions Basics

Chat Completions remains relevant when maintaining an existing messages-based integration. Evaluate Responses for a new workflow or a planned migration rather than rewriting a stable application merely to change terminology.

#Core model

The application supplies role-labelled messages and manages which conversation history is sent. A response can include text or tool-related content; inspect the returned structure rather than assuming every choice is a final answer.

pythonCopy
import os
from openai import OpenAI

client = OpenAI(timeout=20.0)
response = client.chat.completions.create(
    model=os.environ["OPENAI_MODEL"],
    messages=[
        {"role": "system", "content": "Explain programming concepts clearly."},
        {"role": "user", "content": "What is a generator?"},
    ],
)
print(response.choices[0].message.content)

This example assumes the configured model supports Chat Completions.

#Migration considerations

Inventory existing context storage, tools, images, retries and error handling. Responses offers a different item/response model; it is not a field-by-field rename of messages.

Start with one equivalent workflow, compare the results and preserve application-owned permissions and state. Add new tool/agent features deliberately rather than accumulating unsupported assumptions in a legacy wrapper.

#Reference

• Migrating to Responses↗

#Related pages

• OpenAI API overview
• Quickstart
• Function calling