[BUG] The list of models doesn't load in the panel below on the OpenAI Compatible provider

[jwadow]

Plugin Type

VSCode Extension

App Version

4.110.0

Description

I'm switching from Roo Code to Kilo Code, but I'm annoyed by the space-consuming panel at the bottom. Furthermore, the only useful function button (model selection) doesn't work, only showing the current model. The most logical thing, at least for the OpenAI Compatible provider, is to do a regular GET /v1/models cache for this button.

https://github.com/user-attachments/assets/217986e6-fd96-436a-8b20-719d8c9a5813

Reproduction steps

1. Create OpenAI Compatible provider that includes multiple models.
2. Click on "Model" in settings, you can see all your models (also GET /v1/models).
3. Apply the settings and exit to the main workspace.
4. Click on the model in the bottom panel (no GET /v1/models) - it's broken for OpenAI Compatible provider.

Provider

OpenAI Compatible

Model

No response

System Information

No response

[anschmieg]

I am experiencing the same issue (which I believe is different from #1216, which is the only similar one I found) on Kilo Code 4.121.0 with the following config:

{
  "providerProfiles": {
    "currentApiConfigName": "myCustomEndpoint",
    "apiConfigs": {
      "copilot": {
        "rateLimitSeconds": 0,
        "enableReasoningEffort": false,
        "openAiBaseUrl": "https://custom-api.example.com/v1",
        "openAiApiKey": "sk-blank",
        "openAiLegacyFormat": true,
        "openAiModelId": "gpt-4.1",
        "openAiCustomModelInfo": {
          "maxTokens": 64000,
          "contextWindow": 264000,
          "supportsImages": true,
          "supportsPromptCache": false,
          "inputPrice": 0,
          "outputPrice": 0
        },
        "apiProvider": "openai",
        "id": "myCustomEndpoint",
        "models": {
          "gpt-4.1": {
            "id": "gpt-4.1",
            "name": "GPT-4.1"
          },
          "gpt-4o": {
            "id": "gpt-4o",
            "name": "GPT-4o"
          },
          "gpt-4o-mini": {
            "id": "gpt-4o-mini",
            "name": "GPT-4o mini"
          }
        }
      }
    },
    "modeApiConfigs": {
      "architect": "myCustomEndpoint",
      "code": "myCustomEndpoint",
      "ask": "myCustomEndpoint",
      "debug": "myCustomEndpoint",
      "orchestrator": "myCustomEndpoint"
    }
  },
  "globalSettings": {
[...]

The endpoint supports /models listing. Yet, I tried defining the models manually as you can see, which still doesn't do anything, as shown below (image shows other profile name).

[GitCourser]

Indeed, the button below the chat dialog should be "Model Selection" rather than just being able to view the "Model Name" functionality; several people have raised this issue. https://github.com/Kilo-Org/kilocode/issues/3686 https://github.com/Kilo-Org/kilocode/discussions/4037