Vision (Image Understanding)

Vision capabilities are built into LLM models that support multimodal input. Send images alongside text in the Chat Completions API.

How to Use

Pass images as content parts in the messages array:

response = client.chat.completions.create(
    model="gpt-4o",
    messages=[{
        "role": "user",
        "content": [
            {"type": "text", "text": "What's in this image?"},
            {"type": "image_url", "image_url": {"url": "https://example.com/photo.jpg"}}
        ]
    }]
)
print(response.choices[0].message.content)

Image Input Formats

Format	Example
URL	{"type": "image_url", "image_url": {"url": "https://..."}}
Base64	{"type": "image_url", "image_url": {"url": "data:image/png;base64,..."}}

Models with Vision

Model	Max Images	Notes
gpt-4o	20	Best overall vision
claude-sonnet-4	20	Strong document understanding
gemini-2.5-pro	16	1M context for long documents
meta/llama-4-maverick	8	Open-weight alternative

→ Full capability matrix