Vision

Model APIs support text and vision inputs, but multimodal capability depends on the underlying model. Vision-capable models accept images alongside text in the same request, using the OpenAI-compatible image_url content type. The model processes both modalities together, so it can answer questions about image content, compare multiple images, or extract structured data from screenshots. Not all models support vision. Check the table below before sending image inputs.

Supported models

Model	Slug
DeepSeek V4.1 Flash	`deepseek-ai/DeepSeek-V4.1-Flash`
DeepSeek V4.1 Flash Fast	`deepseek-ai/DeepSeek-V4.1-Flash-Fast`
GLM 5.2	`zai-org/GLM-5.2`
GLM 5.2 Fast	`zai-org/GLM-5.2-Fast`
GLM 5.3	`zai-org/GLM-5.3`
GLM 5.3 Fast	`zai-org/GLM-5.3-Fast`
GLM 5.3 Flash	`zai-org/GLM-5.3-Flash`
Kimi K3	`moonshotai/Kimi-K3`
Muse Spark Open	`meta-models/muse-spark-open`
Nemotron 3.5 Super VL	`nvidia/NVIDIA-Nemotron-3.5-Super-VL-120B-A12B`

Nemotron 3.5 Super VL accepts text and image inputs. Audio and video inputs aren’t supported through its Model API.

Send a vision request

Use the image_url content type to include images in your messages. Baseten retrieves image URLs from the inference service, so the URL must be reachable over HTTPS from Baseten’s environment (for example your own object storage, Hugging Face artifact links, or other hosts that allow server-side fetches). Prefer stable, direct HTTPS links. Optional image_url.detail controls preprocessing resolution: low, high, original, or auto (OpenAI-compatible). When in doubt, use auto. Send an image alongside a text prompt like this:

Python
JavaScript
cURL

vision.py

from openai import OpenAI
import os

client = OpenAI(
    base_url="https://inference.baseten.co/v1",
    api_key=os.environ["BASETEN_API_KEY"],
)

response = client.chat.completions.create(
    model="moonshotai/Kimi-K3",
    messages=[
        {
            "role": "user",
            "content": [
                {
                    "type": "text",
                    "text": "Describe the natural environment in the image.",
                },
                {
                    "type": "image_url",
                    "image_url": {
                        "url": "https://huggingface.co/datasets/YiYiXu/testing-images/resolve/main/seashore.png",
                        "detail": "auto",
                    },
                },
            ],
        }
    ],
)

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

vision.js

import OpenAI from "openai";

const client = new OpenAI({
    baseURL: "https://inference.baseten.co/v1",
    apiKey: process.env.BASETEN_API_KEY,
});

const response = await client.chat.completions.create({
    model: "moonshotai/Kimi-K3",
    messages: [
        {
            role: "user",
            content: [
                {
                    type: "text",
                    text: "Describe the natural environment in the image.",
                },
                {
                    type: "image_url",
                    image_url: {
                        url: "https://huggingface.co/datasets/YiYiXu/testing-images/resolve/main/seashore.png",
                        detail: "auto",
                    },
                },
            ],
        },
    ],
});

console.log(response.choices[0].message.content);

Request

curl https://inference.baseten.co/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $BASETEN_API_KEY" \
  -d '{
    "model": "moonshotai/Kimi-K3",
    "messages": [
      {
        "role": "user",
        "content": [
          {
            "type": "text",
            "text": "Describe the natural environment in the image."
          },
          {
            "type": "image_url",
            "image_url": {
              "url": "https://huggingface.co/datasets/YiYiXu/testing-images/resolve/main/seashore.png",
              "detail": "auto"
            }
          }
        ]
      }
    ]
  }'

Image and video limits

Baseten’s media encoder enforces the following per-request limits:

Limit	DeepSeek V4.1 Flash	DeepSeek V4.1 Flash Fast	Kimi K3	Muse Spark Open	GLM 5.2	GLM 5.2 Fast	GLM 5.3	GLM 5.3 Fast	GLM 5.3 Flash
Max images per request	64	16	96	8	8	8	8	8	64
Max videos per request	Not supported	Not supported	12	2	Not supported	Not supported	Not supported	Not supported	Not supported
Max total media size per request (URL)	200 MB	64 MB	320 MB	200 MB	100 MB	100 MB	100 MB	100 MB	100 MB
Max size per image (URL)	200 MB	32 MB	80 MB	10 MB	80 MB	80 MB	80 MB	80 MB	80 MB
Max size per video (URL)	Not supported	Not supported	150 MB	200 MB	Not supported	Not supported	Not supported	Not supported	Not supported
Max video duration	Not supported	Not supported	30 seconds	300 seconds	Not supported	Not supported	Not supported	Not supported	Not supported
Max request body (base64)	45 MB	64 MB	62 MB	100 MiB	62 MB	62 MB	62 MB	62 MB	100 MB

Pass media by URL whenever you can. The vision encoder fetches each URL and enforces the per-file and total-media caps directly, while the request body stays small. Base64-encoded media travels inside the request body and counts toward the model’s request-body limit. Muse Spark Open accepts JPEG, PNG, WebP, GIF, BMP, and TIFF images. For animated GIFs, it uses only the first frame. It accepts MP4, MOV, MKV, and WebM videos with H.264, H.265, VP9, or AV1 encoding, up to 4K at 60 frames per second. It samples at most 96 frames per video and ignores the audio track. Send speech as a separate audio_url part. Muse Spark Open enforces a 200 MB combined limit for images, videos, and audio, plus a 896 MiB encoded-media budget per request. A request can fit within the file-size limits and still exceed the encoded-media budget. Combining 8 large images, 2 five-minute videos, and 4 five-minute audio files can return HTTP 400. Split large combinations across requests. Send files larger than about 75 MB by URL. Base64 encoding increases their size by about a third, and the complete request must fit within the 100 MiB request-body limit. Muse Spark Open doesn’t accept PDF files. Extract their text before sending it.

Pricing

There is no additional per-image fee. Images are converted to input tokens and priced at the model’s standard input rate. Higher resolution images produce more tokens and cost more to process. The exact conversion from pixels to tokens depends on the model. Kimi K3 divides each image into 28×28 pixel tiles where each tile becomes one input token. The cost table below uses Kimi K3’s uncached input rate ($3.00 per million tokens); for other models, use the rates on the Model APIs pricing page.

Image resolution	Tiles	Input tokens	Cost at $3.00/M
256×256	100	100	$0.0003
512×512	361	361	$0.0011
1024×1024	1,369	1,369	$0.0041
1920×1080	2,691	2,691	$0.0081

For videos, token count scales with both resolution and the number of sampled frames.

Next steps

Model APIs overview

Supported models, pricing, and the feature support matrix

Chat Completions reference

Full request and response schema for the image_url content type

Get started

Model APIs

Inference

Training

Sandboxes

Development

Deployment

Engines

Model Labs

Organization

Hosting options

Observability

Troubleshooting

Supported models

Send a vision request

Image and video limits

Pricing

Next steps

Model APIs overview

Chat Completions reference

​Supported models

​Send a vision request

​Image and video limits

​Pricing

​Next steps

Model APIs overview

Chat Completions reference

Supported models

Send a vision request

Image and video limits

Pricing

Next steps