Skip to main content
Chat Completions API is a compatibility surface. For new LLM integrations, prefer Responses API.
Use this API when you already have code built around OpenAI-style messages[], choices[0].message, or chat-compatible SDK behavior. If you are starting fresh, use Responses API instead.

​
Best Fit

  • existing OpenAI SDK integrations
  • chat UIs already built around messages[]
  • systems that expect tool_calls, response_format, and chunked chat streaming

​
When to stay on Chat Completions

  • your app already uses OpenAI chat-shaped requests everywhere
  • your client code expects choices[0].message
  • the migration cost is higher than the near-term benefit

​
Request Model

Common fields include:
  • model
  • messages
  • tools
  • response_format
  • stream
  • stream_options
  • reasoning_effort
  • web_search_options

​
Quick Example

from openai import OpenAI

client = OpenAI(
    base_url="https://api.naga.ac/v1",
    api_key="YOUR_API_KEY",
)

completion = client.chat.completions.create(
    model="gpt-4o-mini",
    messages=[
        {"role": "user", "content": "Explain retries in one paragraph."}
    ],
)

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

​
Response Model

Successful non-streaming responses follow the OpenAI chat shape. 
{
  "object": "chat.completion",
  "choices": [
    {
      "message": {
        "role": "assistant",
        "content": "Retries improve resilience when failures are transient."
      },
      "finish_reason": "stop"
    }
  ]
}

​
Common mistakes

  • starting a new integration on Chat Completions when Responses would be simpler
  • assuming every capability maps 1:1 to Responses item semantics
  • forgetting that this is a compatibility layer over the same platform catalog

​
Learn The API In Detail

Streaming

Stream chat chunks and handle terminal chunks correctly.

Tool Calling

Keep OpenAI-style chat tool workflows.

Reasoning

Control reasoning behavior on reasoning-capable models.

Structured Outputs

Enforce JSON or schema-shaped outputs.

Multimodal Content

Send text, images, files, and audio in chat content blocks.

Web Search

Enable grounded answers with search-aware models.

Migration to Responses

Move from chat-shaped requests to the newer Responses model.

​
Reference