Error codes Models

PULSE-MODEL-004 The model did not accept the request

What does this mean

The model server received the request and responded that it would not execute it: the request, according to it, was incorrect. The agent's progress is stopped; what the agent managed to do before remains in the working copy of the task.

Why does this happen

The server responds with code 400, 404, 413 or 422. Most often:

  • the model cannot call tools (tools), and the agent needs them - this happens with small local models;
  • there is no such model on the server: it was deleted or the name was misspelled - Pulse writes “model not found on the server”;
  • The session history does not fit into the model context window: context_length_exceeded, prompt is too long, "request too big";
  • The server does not understand the request format: invalid_request_error, for GigaChat - “GigaChat did not accept the request.”

In “More details” - the server’s entire response: it shows which case is yours.

What to do

  1. Open "More details" and read the server response.
  2. A model without tools or a model is not on the server - select another one in the “Model” chip (for a local model, the list is available using the “Check” button in “Settings → Models”).
  3. The history is too long - start a new task session or increase the model context window in its settings.
  4. Repeating the same turn with the "Repeat turn" button usually gets the same response - repeat after changing the model or settings.

When to write to support

Write to us if the model knows how to call tools, but Pulse still receives this response. Click “Send report” in the error card - a report form will open, where the code and server response have already been entered, the secrets are hidden in them. Add which server and which model you chose, and click “View and Send”: the window will show exactly what is left - check and click “Submit”. If the cloud is turned off, the button is called “View and save” - send the resulting file to us in any way.