feat: add hide "base" models flag to model spec

[aron-muon]

Summary

Adds a new hideBaseModels configuration option to model specs that allows hiding base/raw models from the model selector when model specs are defined for an endpoint.

Related Discussion: #10220 (reply in thread)
Documentation PR: LibreChat-AI/librechat.ai#465

Problem

When using modelSpecs to provide preconfigured models (e.g., a "gpt-5-codex" with specific settings like Responses API enabled), both the model spec AND the base model appear in the selector. This creates confusion:

• Users see duplicate entries (the spec and the raw model)
• No clear indication of which to choose
• Adds unnecessary clutter when administrators want to present only curated model options

Solution

This PR implements the hideBaseModels option suggested in the discussion:

modelSpecs:
  hideBaseModels: true  # Hide base models when specs exist for an endpoint
  list:
    - name: "My Custom Model"
      # ...

When hideBaseModels: true is set, base models are hidden only for endpoints that have model specs defined. Endpoints without any specs will continue to show their models normally.

Change Type

• New feature (non-breaking change which adds functionality)

Implementation

• Added hideBaseModels boolean to the specsConfigSchema (defaults to false)
• Exposed the value from startup config via the model selector context
• Updated the endpoint item rendering logic to conditionally hide base models when hideBaseModels is true and the endpoint has specs defined
• Added unit tests for schema validation
• Added example configuration to librechat.example.yaml

Testing

Unit Tests

npx jest specs/models.spec.ts

Tests verify:

• hideBaseModels defaults to false when not provided
• hideBaseModels: true is accepted
• hideBaseModels: false is accepted
• Non-boolean values are rejected

Manual Testing

1. Set hideBaseModels: false (or omit) → Base models show as normal alongside specs
2. Set hideBaseModels: true with specs for an endpoint → Only specs are shown, base models hidden
3. Set hideBaseModels: true for endpoint with NO specs → Base models still shown (no specs to replace them)

Checklist

• My code adheres to this project's style guidelines
• I have performed a self-review of my own code
• I have commented in any complex areas of my code
• I have made pertinent documentation changes
• My changes do not introduce new warnings
• I have written tests demonstrating that my changes are effective or that my feature works
• Local unit tests pass with my changes

[aron-muon]

@danny-avila this is ready for review!

[aron-muon]

With a config like this,

  "modelSpecs": {
    "enforce": false,
    "prioritize": true,
    "hideBaseModels": true,
    "addedEndpoints": [
      "anthropic",
      "google",
      "agents"
    ],
    "list": [
      {
        "name": "claude-opus-4-5",
        "label": "Claude Opus 4.5",
        "iconURL": "anthropic",
        "description": "Latest slow flagship model with Claude.",
        "group": "anthropic",
        "showIconInMenu": true,
        "showIconInHeader": true,
        "preset": {
          "endpoint": "anthropic",
          "model": "claude-opus-4-5",
          "web_search": true,
          "promptCache": true,
          "modelLabel": "Claude Opus 4.5"
        }
      },
      {
        "name": "claude-opus-4-1",
        "label": "Claude Opus 4.1",
        "iconURL": "anthropic",
        "description": "Slow, but accurate chat with Claude.",
        "group": "anthropic",
        "showIconInMenu": true,
        "showIconInHeader": true,
        "preset": {
          "endpoint": "anthropic",
          "model": "claude-opus-4-1",
          "web_search": true,
          "promptCache": true,
          "modelLabel": "Claude Opus 4.1"
        }
      },
      {
        "name": "claude-sonnet-4-5",
        "label": "Claude Sonnet 4.5",
        "iconURL": "anthropic",
        "description": "Medium-fast chat with Claude.",
        "group": "anthropic",
        "showIconInMenu": true,
        "showIconInHeader": true,
        "preset": {
          "endpoint": "anthropic",
          "model": "claude-sonnet-4-5",
          "web_search": true,
          "promptCache": true,
          "modelLabel": "Claude Sonnet 4.5"
        }
      },
      {
        "name": "claude-haiku-4-5",
        "label": "Claude Haiku 4.5",
        "iconURL": "anthropic",
        "description": "Fast chat with Claude.",
        "group": "anthropic",
        "showIconInMenu": true,
        "showIconInHeader": true,
        "preset": {
          "endpoint": "anthropic",
          "model": "claude-haiku-4-5",
          "web_search": true,
          "promptCache": true,
          "modelLabel": "Claude Haiku 4.5"
        }
      }
    ]
  },

Without this new feature, this is shown:

But with this new feature, this is shown instead - much less confusing, much more understandable.

[MaximilianoGutierrez]

I used this configuration to hide the base models; however, they are still appearing despite hideBaseModels being set to true. The expectation was that only the models explicitly defined in the list section would be visible.

modelSpecs:
  enforce: true
  prioritize: true
  hideBaseModels: true
  #addedEndpoints: ["azureOpenAI", "google", "anthropic"]
  addedEndpoints:
    - azureOpenAI
    - google
    - anthropic
  list:
  #Anthropic models
    - name: "claude-sonnet-4-5"
      label: "claude-sonnet-4-5"
      default: true
      iconURL: "anthropic"
      description: "Prueba de descripcion"
      group: "anthropic"
      showIconInMenu: true
      showIconInHeader: true
      preset:
        endpoint: "anthropic"
        model: "claude-sonnet-4-5"
        #useResponsesApi: true
        promptPrefix: |
          You are claude-sonnet-4-5, a helpful AI assistant created by Anthropic.
          The current date is {{current_date}}.
          IMPORTANT: Calculate the day of the week from the date above. Do not assume or guess.