Image input

Put an image_url part next to the text in a message's content array. The URL is either a public https:// link or a base64 data: URL holding the file itself.

curl https://api.sara-ai.kz/v1/chat/completions \
  -H "Authorization: Bearer $SARA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gemma-4-31b",
    "messages": [{
      "role": "user",
      "content": [
        {"type": "text", "text": "What does this image say?"},
        {"type": "image_url", "image_url": {"url": "https://sara-ai.kz/en/opengraph-image"}}
      ]
    }]
  }'

Which models accept images

• gemma-4-31b accepts images.
• glm-5.2-744b is text-only. A request that sends it an image returns 404.

Good to know

• PNG and JPEG are supported.
• A message can carry several images: add one image_url part per image.
• An image is billed as input tokens: about 260–280 per image, whatever its dimensions.
• Images work with streaming, thinking and structured outputs.
• A request can be up to 32 MB. A base64 image is about a third larger than the file it encodes.
• /v1/messages and /v1/responses accept images too, each in its own format.