FIM not working with OpenRouter

[guillaumevpayet]

Describe the bug
When using OpenRouter for FIM, nothing happens. Looking at the logs (below), my theory is the model field in the request body is missing.

To Reproduce
Create an OpenRouter FIM provider with:

• Type: fim
• FIM Template: automatic
• Provider: openrouter
• Protocol: https
• Model Name: qwen/qwen3-coder
• Hostname: openrouter.ai
• API Path: /api/v1/completions
• API Key:

Expected behavior
FIM to work

Logging

[INFO] ***Twinny Stream Debug***
    Streaming response from openrouter.ai.
    Request body:
    {
  "prompt": "<PRE> \n\n# Language: Perl (perl) \n# File uri: untitled:Untitled-1 (perl) \nStart of my code <SUF> \nEnd of my code <MID>",
  "stream": true,
  "temperature": 0.2,
  "n_predict": 512
}

    Request options:
    {
  "hostname": "openrouter.ai",
  "path": "/api/v1/completions",
  "protocol": "https",
  "method": "POST",
  "headers": {
    "Content-Type": "application/json",
    "Authorization": "Bearer <redacted>"
  }
}

    Number characters in all messages = 0
[ERROR_error] Fetch error
  Error Type: Error
  Error Message: Server responded with status code: 400

API Provider
OpenRouter

Chat or Auto Complete?
It is an issue with FIM. Using the same model and OpenRouter works fine for chatting.

Model Name
qwen/qwen3-coder (but it will be the same with other models)

Desktop:

• OS: Linux/Windows (tried on both)
• Browser: Firefox
• Version: 142

Additional context
As per the OpenRouter API documentation, the model field is required: https://openrouter.ai/docs/api-reference/completion