Better model picker & model reasoning controls

[Blaumaus]

Summary by CodeRabbit

• New Features

  • Reasoning effort selector (Off / Low / Medium / High) for compatible models — included with message submissions and used to influence AI responses.
  • Model lists and settings surface richer metadata: version labels, price tiers, and feature icons (e.g., vision, reasoning).
  • Chat creation and inline edit flows include reasoning effort when present.

• Bug Fixes

  • Invalid/unsupported reasoning selections are automatically cleared when a model no longer supports them.

[coderabbitai]

Caution

Review failed

The pull request is closed.

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro Plus

Run ID: 861fa196-20e3-4d20-b249-23987d9e6cc8

📥 Commits

Reviewing files that changed from the base of the PR and between c2aabf8 and 4e500a4.

📒 Files selected for processing (1)

• packages/database/src/model-registry.ts

---

📝 Walkthrough

Walkthrough

Adds reasoning-effort support across stack: new reasoning types and registry, enriched model metadata, UI selector and form wiring, route validation, job payload plumbing, and worker/provider streaming options for reasoning efforts.

Changes

Reasoning Effort System

Layer / File(s)	Summary
Type Definitions & Exports packages/database/src/types.ts, packages/database/src/index.ts	Adds REASONING_LEVELS, ReasoningLevel, isReasoningLevel; extends ChatJobData with optional reasoningEffort; re-exports new types and helpers.
Model Registry & Metadata packages/database/src/model-registry.ts	Introduces MODEL_REGISTRY with ModelMetadata (provider, display, price, features, optional reasoningLevels, context); adds getModelMetadata and getModelMetadataMatch with delimiter-aware longest-prefix matching.
Model Types & Enrichment apps/web/app/lib/models.ts, apps/web/app/lib/models.server.ts	ModelInfo gains versionLabel, knowledgeCutoff, priceTier, features, reasoningLevels; enrichCachedModel populates metadata; getCachedModels/getEnabledModels return enriched ModelInfo and disambiguate variant labels.
UI: Chat Input & Model Options apps/web/app/components/ChatInput.tsx	Adds reasoningEffort and onReasoningEffortChange props; conditionally renders ReasoningSelector when model exposes reasoningLevels and handler is provided; clears unsupported selection on model change; includes reasoningEffort as hidden form input; refactors model option display, REASONING_LABELS, FEATURE_ICONS, and theme classes.
Routes: Chat Actions & New Chat apps/web/app/routes/_chat.chat.$chatId.tsx, apps/web/app/routes/_chat.chat._index.tsx, apps/web/app/routes/api.chat.send.tsx	Reads reasoningEffort from form data, validates via isReasoningLevel, and includes validated value in addChatJob for new messages, edits, retries, and API send. Client state forwards reasoningEffort into ChatInput. Model <option> labels show versionLabel when present.
Settings UI apps/web/app/routes/settings.models.tsx	ModelRow displays versionLabel, priceTier via PriceTierBadge, feature icons from features, and joined reasoningLevels; uses full cached model shape.
Worker: Streaming & Processing apps/worker/src/utils.ts, apps/worker/src/processors/chat.ts	streamAIResponse signature adds reasoningEffort; computes providerOptions via buildProviderOptions using model metadata and reasoningEffort, and passes options into streamText; processChatJob forwards reasoningEffort to streaming. TITLE_MODELS adjusted.
Worker: Model Refresh apps/worker/src/processors/refresh.ts	Google model refresh now filters by isGoogleChatModelId (explicit modality exclusions) instead of substring matching.
Model ID Helpers apps/worker/src/utils.ts	isOpenAIModelId and isGoogleChatModelId refined to exclude non-chat modalities (audio/image/realtime/tts/etc.).

sequenceDiagram
    participant User as User
    participant UI as ChatInput
    participant Route as Server Route
    participant Queue as Job Queue
    participant Worker as Worker
    participant Provider as LLM Provider

    User->>UI: choose reasoning level & submit message
    UI->>Route: POST message + reasoningEffort
    Route->>Route: validate with isReasoningLevel
    Route->>Queue: addChatJob({...reasoningEffort})
    Queue->>Worker: deliver job
    Worker->>Worker: getModelMetadata(modelId)
    Worker->>Worker: buildProviderOptions(modelId, reasoningEffort)
    Worker->>Provider: streamText(..., providerOptions)
    Provider-->>Worker: stream tokens
    Worker-->>UI: stream response updates

Loading

🎯 4 (Complex) | ⏱️ ~45 minutes

• ChathouseHQ/chathouse#13: Implements the same reasoningEffort support across UI, types/registry, routes, worker utils/processors, and model metadata.

"🐰
I nibble on metadata bright,
Levels set for thinking's light,
From selector to stream in flight,
Tokens dance with tidy might,
Hopping, reason wakes the night."

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 3.23% which is insufficient. The required threshold is 80.00%.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The pull request title accurately summarizes the primary changes: enhanced model picker UI with improved display of model metadata, and addition of reasoning effort controls throughout the application.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

📝 Generate docstrings

• Create stacked PR
• Commit on current branch

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch feature/better-model-picker

---

_{Comment @coderabbitai help to get the list of available commands and usage tips.}