Models supporting image generation are not detected

[Tracker]
Update Summary: Models supporting image generation are not detected
Short Description: Models supporting image generation are not detected
Check-in Date: 12/16/2025
Metadata is used by the AI Tracker. Docs and additional fields here.
[/Tracker]

In my LLM instance, when using model/info we can see several keys indicating the operation supported by the models

"mode": "image_generation",
"supports_system_messages": null,
"supports_response_schema": null,
"supports_vision": null,
"supports_function_calling": null,
"supports_tool_choice": null,
"supports_assistant_prefill": null,
"supports_prompt_caching": null,
"supports_audio_input": null,
"supports_audio_output": null,
"supports_pdf_input": null,
"supports_embedding_image_input": null,
"supports_native_streaming": null,
"supports_web_search": null,
"supports_url_context": null,
"supports_reasoning": null,
"supports_computer_use": null,

Actual code detects imageInput and imageOutput differently than than the output of the API

 public static function createFromResponse(\stdClass $response): self {
    $model_info = $response->model_info;
    return new self(
      $response->model_name,
      $model_info->supports_image_input ?? FALSE,
      $model_info->supports_image_output ?? FALSE,
      $model_info->supports_audio_input ?? FALSE,
      $model_info->supports_audio_output ?? FALSE,
      $model_info->supports_video_output ?? FALSE,
      ($model_info->mode ?? NULL) === 'embedding',
      ($model_info->mode ?? NULL) === 'chat',
      $model_info->supports_moderation ?? FALSE,
      $model_info->supported_openai_params ?? [],
    );
  }

- supports_image_input is under supports_vision
- supports_image_output is under mode: "image_generation"