Fixing "Model does not accept images"

This 400 means the request carried an image and the chosen model only reads text. Retrying the same request will fail the same way every time.

What the error says

HTTP 400
{
  "error": {
    "message": "Model \"glm-5.3-outlaw\" does not accept images. Remove the images from the request, or use a model that supports vision.",
    "type": "invalid_request_error",
    "code": "images_unsupported",
    "param": "messages"
  }
}

On /v1/messages the same message comes back as a 400 invalid_request_error. The model id in the message is the one you sent, so you can see at once which model refused.

Which models read images

glm-5.3-outlaw is text only. The others in the line accept images. The supports_vision field in GET /v1/models says so for each model, and the models page shows the same.

Most API failures from the model are reported as a short outage, but this one is not, on purpose: a client that sees "try again" would resend the image forever.

Where the image comes from in SillyTavern

Inline images. An image attached to a message, or the option to send inline images, puts an image part in the request.
Earlier messages. An image anywhere in the chat history is sent again with every new message, so the error keeps coming after you stop attaching.
Image captioning or an extension set to use the chat model for captions.

Step by step fix

Decide whether you need images. If yes, switch to a model with vision, such as outlaw-1 or glm-5.3-flash-outlaw.
If you want to keep glm-5.3-outlaw, turn off sending inline images in SillyTavern.
Delete or hide the messages in the chat that contain images, or start a new chat, so history stops carrying them.
In code, remove image_url content parts from messages before sending to a text-only model.
Retry.

Telling it apart from similar errors

A 503 Model overloaded is a transient failure and worth retrying; see 503. A 400 model_not_found means the id itself is wrong; see Unknown model. A 400 about context length is size, not content; see context length exceeded.

FAQ

Am I charged for the refused request?

The model refuses before generating anything, so no output is produced or billed.

Can the API strip images for me?

No. It passes the request through as sent. Remove images on the client side or pick a vision model.

Fixing "Unknown model" (model_not_found)This error means the model field held an id the API does not serve. The key and balance were fine; the name was not. Fixing "maximum context length" (context_length_exceeded)This 400 means the prompt plus the reply you reserved add up to more tokens than the model can hold. It is about the request, not an outage, so retrying unchanged will fail again. Fixing 503 "Model overloaded. Please try again in a minute."A 503 means the model behind the API failed to answer, and the API had already retried before telling you. It is usually brief.