Visual Understanding

The LongCat-2.5-Preview model accepts images and videos alongside text. You can ask the model to describe an image, read text from a screenshot, analyze a chart, or understand video content.

Passing Images

There are two ways to provide images. Both use the standard OpenAI-compatible chat completions format, where content is an array of blocks rather than a plain string.

Method 1: Base64-Encoded Image (Inline)

Encode the image and embed it directly in the request as a data: URL — the simplest option for local files. Base64 data counts toward the request body size limit (see Limits).

import base64
from openai import OpenAI

client = OpenAI(
    api_key="YOUR_API_KEY",
    base_url="https://api.longcat.chat/openai"
)

with open("image.jpg", "rb") as f:
    b64 = base64.b64encode(f.read()).decode("utf-8")

response = client.chat.completions.create(
    model="LongCat-2.5-Preview",
    messages=[
        {
            "role": "user",
            "content": [
                {"type": "text", "text": "What is in this image?"},
                {
                    "type": "image_url",
                    "image_url": {"url": f"data:image/jpeg;base64,{b64}"},
                },
            ],
        }
    ],
)
print(response.choices[0].message.content)

curl https://api.longcat.chat/openai/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -d '{
    "model": "LongCat-2.5-Preview",
    "messages": [
      {
        "role": "user",
        "content": [
          {"type": "text", "text": "What is in this image?"},
          {"type": "image_url", "image_url": {"url": "data:image/jpeg;base64,<BASE64_DATA>"}}
        ]
      }
    ]
  }'

Method 2: External Image URL

Pass a publicly accessible http(s) link, and the model downloads the image automatically.

response = client.chat.completions.create(
    model="LongCat-2.5-Preview",
    messages=[
        {
            "role": "user",
            "content": [
                {"type": "text", "text": "Describe this image."},
                {
                    "type": "image_url",
                    "image_url": {"url": "https://example.com/image.jpg"},
                },
            ],
        }
    ],
)
print(response.choices[0].message.content)

Passing Videos

Videos can only be passed via URL. The model downloads and parses the video content automatically. Videos are carried in video_url blocks within the same content block array:

response = client.chat.completions.create(
    model="LongCat-2.5-Preview",
    messages=[
        {
            "role": "user",
            "content": [
                {"type": "text", "text": "What happens in this video?"},
                {
                    "type": "video_url",
                    "video_url": {"url": "https://example.com/video.mp4"},
                },
            ],
        }
    ],
)
print(response.choices[0].message.content)

Token Usage

Images and videos are converted to tokens based on their dimensions and billed together with text tokens.

Estimation formula: N_image = total image pixels ÷ 784 (each visual token covers 28 × 28 pixels). For actual PromptTokens accounting, add the two special tokens <img_start> and <img_end>, so the PromptTokens increase per image ≈ N_image + 2.

You can also use the calculator below to estimate the token cost of a single image:

Image Token Calculator

Formula: N = (W × H) ÷ 784

Width (px)×Height (px)

Estimated visual tokens--

Estimate only. Actual token count is subject to the usage field in the API response.

Limits

Image Limits

Item	Value
Supported formats	PNG, JPG, JPEG, WebP, GIF (SVG, HEIC/HEIF not supported)
Input methods	Base64 encoding, URL
Original aspect ratio	Over 200:1 returns an error
Max size per image	10 MB
Max images per request	50 (tentative)

Video Limits (tentative)

Item	Value
Supported formats	MP4, MOV, MKV, AVI, WebM
Input method	URL only
Max size per video	50 MB
Max duration	1 hour

API Information

OpenAI-Compatible API

Image and video understanding uses the standard OpenAI-compatible chat completions endpoint:

• Request path: POST https://api.longcat.chat/openai/v1/chat/completions
• Model name: LongCat-2.5-Preview
• Image content block: {"type": "image_url", "image_url": {"url": "..."}} — url accepts a public URL or a data: URL (Base64)
• Video content block: {"type": "video_url", "video_url": {"url": "..."}} — url accepts public URLs only

Anthropic-Compatible API

Besides the OpenAI-compatible endpoint, images can also be passed via the Anthropic-compatible /messages endpoint (base_url: https://api.longcat.chat/anthropic). The difference lies in the image content block structure: Anthropic uses an image block whose source object has a type of either base64 or url:

import anthropic

client = anthropic.Anthropic(
    api_key="YOUR_API_KEY",
    base_url="https://api.longcat.chat/anthropic"
)

message = client.messages.create(
    model="LongCat-2.5-Preview",
    max_tokens=1024,
    messages=[
        {
            "role": "user",
            "content": [
                {"type": "text", "text": "What is in this image?"},
                {
                    "type": "image",
                    "source": {
                        "type": "base64",
                        "media_type": "image/jpeg",
                        "data": "<BASE64_DATA>",
                    },
                },
            ],
        }
    ],
)
print(message.content)