Skip to main content
Chat Completions API accepts either plain string message content or arrays of typed content blocks. Use this when you already depend on chat-style messages[] but still need images, PDFs, or audio in the same request.

​
Supported Content Blocks

Block typeShapeNotes
text{ "type": "text", "text": "..." }Plain text
image_url{ "type": "image_url", "image_url": { "url": "...", "detail": "auto" } }url accepts http, https, or data; detail can be low, high, or auto
file{ "type": "file", "file": { "filename": "...", "file_data": "..." } }file_data accepts http, https, or data
input_audio{ "type": "input_audio", "input_audio": { "data": "...", "format": "wav" } }data must be raw base64, not a data URI; common formats include wav and mp3

​
When to use this surface

  • you already have chat-based client code and want to keep it
  • you need multimodal inputs but do not want to migrate to Responses yet
If you are starting from scratch, Responses API is usually the cleaner multimodal path.

​
Image 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": [
                {"type": "text", "text": "Describe this image briefly."},
                {
                    "type": "image_url",
                    "image_url": {
                        "url": "https://example.com/receipt.png",
                        "detail": "auto",
                    },
                },
            ],
        }
    ],
)

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

​
File 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-5",
    messages=[
        {
            "role": "user",
            "content": [
                {"type": "text", "text": "Summarize this PDF."},
                {
                    "type": "file",
                    "file": {
                        "filename": "report.pdf",
                        "file_data": "https://example.com/report.pdf",
                    },
                },
            ],
        }
    ],
)

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

​
Audio 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="gemini-2.5-flash",
    messages=[
        {
            "role": "user",
            "content": [
                {"type": "text", "text": "Tell me what is said in this audio."},
                {
                    "type": "input_audio",
                    "input_audio": {
                        "data": "BASE64_AUDIO",
                        "format": "wav",
                    },
                },
            ],
        }
    ],
)

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

​
Caveats

  • input_audio.data must be raw base64, not a data: URL
  • multimodal support still depends on the selected model
  • use the direct Audio API for transcription, translation, and text-to-speech
  • use the direct Images API for primary image generation and image edit workflows

​
Common mistakes

  • mixing chat content-block syntax with Responses input-part syntax
  • assuming every model supports every multimodal block type
  • using this surface for dedicated audio or image workflows that belong on the direct APIs