feat(ai): refresh AI models with deprecation support and multi-provider defaults (BREAKING: deploy server before frontend please) (#16503)

## Summary - Add latest AI models from OpenAI (GPT-4.1, o3, o4-mini), Anthropic (Claude 4.5 Opus/Sonnet/Haiku), and xAI (Grok 4.1) - Mark deprecated models (GPT-4o, GPT-4o-mini, GPT-4-turbo, Claude Opus 4, Claude Sonnet 4) with a `deprecated` flag - Split AI models into separate files per provider for better maintainability - Support comma-separated default model lists for automatic fallback across providers (works out of the box for self-hosters regardless of which provider they configure) - Filter deprecated models from dropdown selection while keeping them functional for existing agents ## Changes ### New Models Added | Provider | Models | |----------|--------| | OpenAI | gpt-4.1, gpt-4.1-mini, o3, o4-mini | | Anthropic | claude-opus-4-5, claude-sonnet-4-5, claude-haiku-4-5 | | xAI | grok-4-1-fast-reasoning | ### Deprecated Models - gpt-4o, gpt-4o-mini, gpt-4-turbo (OpenAI) - claude-opus-4-20250514, claude-sonnet-4-20250514 (Anthropic) ### Config Changes Default model configs now support comma-separated fallback lists: - `DEFAULT_AI_SPEED_MODEL_ID=gpt-4.1-mini,claude-haiku-4-5-20251001,grok-3-mini` - `DEFAULT_AI_PERFORMANCE_MODEL_ID=gpt-4.1,claude-sonnet-4-5-20250929,grok-4` ## Test plan - [x] Unit tests pass - [x] Typecheck passes - [x] Lint passes - [ ] Verify deprecated models don't appear in model dropdowns - [ ] Verify agents with deprecated models still work correctly - [ ] Verify default model fallback works when only one provider is configured
2025-12-11 21:47:38 +01:00
parent 999bc84b17
commit a13727335b
15 changed files with 655 additions and 264 deletions
@@ -0,0 +1,64 @@
+export enum ModelProvider {
+  NONE = 'none',
+  OPENAI = 'openai',
+  ANTHROPIC = 'anthropic',
+  OPENAI_COMPATIBLE = 'open_ai_compatible',
+  XAI = 'xai',
+}
+
+export const DEFAULT_FAST_MODEL = 'default-fast-model' as const;
+export const DEFAULT_SMART_MODEL = 'default-smart-model' as const;
+
+export type ModelId =
+  | typeof DEFAULT_FAST_MODEL
+  | typeof DEFAULT_SMART_MODEL
+  // OpenAI models
+  | 'gpt-4o'
+  | 'gpt-4o-mini'
+  | 'gpt-4-turbo'
+  | 'gpt-4.1'
+  | 'gpt-4.1-mini'
+  | 'o3'
+  | 'o4-mini'
+  // Anthropic models
+  | 'claude-opus-4-20250514'
+  | 'claude-sonnet-4-20250514'
+  | 'claude-3-5-haiku-20241022'
+  | 'claude-opus-4-5-20251101'
+  | 'claude-sonnet-4-5-20250929'
+  | 'claude-haiku-4-5-20251001'
+  // xAI models
+  | 'grok-3'
+  | 'grok-3-mini'
+  | 'grok-4'
+  | 'grok-4-1-fast-reasoning'
+  | string; // Allow custom model names
+
+export type SupportedFileType =
+  | 'image/png'
+  | 'image/jpeg'
+  | 'image/gif'
+  | 'image/webp'
+  | 'application/pdf'
+  | 'text/plain'
+  | 'text/html'
+  | 'text/csv'
+  | 'application/json';
+
+export interface AIModelConfig {
+  modelId: ModelId;
+  label: string;
+  description: string;
+  provider: ModelProvider;
+  inputCostPer1kTokensInCents: number;
+  outputCostPer1kTokensInCents: number;
+  contextWindowTokens: number;
+  maxOutputTokens: number;
+  supportedFileTypes?: SupportedFileType[];
+  doesSupportThinking?: boolean;
+  nativeCapabilities?: {
+    webSearch?: boolean;
+    twitterSearch?: boolean;
+  };
+  deprecated?: boolean;
+}
@@ -10,19 +10,56 @@ import {
 } from './ai-models.const';

 describe('AI_MODELS', () => {
-  it('should contain all expected models', () => {
-    expect(AI_MODELS).toHaveLength(9);
-    expect(AI_MODELS.map((model) => model.modelId)).toEqual([
-      'gpt-4o',
-      'gpt-4o-mini',
-      'gpt-4-turbo',
-      'claude-opus-4-20250514',
-      'claude-sonnet-4-20250514',
-      'claude-3-5-haiku-20241022',
-      'grok-3',
-      'grok-3-mini',
-      'grok-4',
-    ]);
+  it('should have at least one model per provider', () => {
+    const providers = [
+      ModelProvider.OPENAI,
+      ModelProvider.ANTHROPIC,
+      ModelProvider.XAI,
+    ];
+
+    providers.forEach((provider) => {
+      const modelsForProvider = AI_MODELS.filter(
+        (model) => model.provider === provider,
+      );
+
+      expect(modelsForProvider.length).toBeGreaterThan(0);
+    });
+  });
+
+  it('should have all required fields for each model', () => {
+    AI_MODELS.forEach((model) => {
+      expect(model.modelId).toBeDefined();
+      expect(model.label).toBeDefined();
+      expect(model.description).toBeDefined();
+      expect(model.provider).toBeDefined();
+      expect(model.inputCostPer1kTokensInCents).toBeDefined();
+      expect(model.outputCostPer1kTokensInCents).toBeDefined();
+      expect(model.contextWindowTokens).toBeGreaterThan(0);
+      expect(model.maxOutputTokens).toBeGreaterThan(0);
+    });
+  });
+
+  it('should have unique model IDs', () => {
+    const modelIds = AI_MODELS.map((model) => model.modelId);
+    const uniqueModelIds = new Set(modelIds);
+
+    expect(uniqueModelIds.size).toBe(modelIds.length);
+  });
+
+  it('should have at least one non-deprecated model per provider', () => {
+    const providers = [
+      ModelProvider.OPENAI,
+      ModelProvider.ANTHROPIC,
+      ModelProvider.XAI,
+    ];
+
+    providers.forEach((provider) => {
+      const activeModelsForProvider = AI_MODELS.filter(
+        (model) => model.provider === provider && !model.deprecated,
+      );
+
+      expect(activeModelsForProvider.length).toBeGreaterThan(0);
+    });
  });
 });

@@ -140,4 +177,50 @@ describe('AiModelRegistryService', () => {
      'Model with ID non-existent-model not found',
    );
  });
+
+  it('should find first available model from comma-separated list', () => {
+    // First model not available, second model available
+    MOCK_CONFIG_SERVICE.get.mockReturnValue(
+      'gpt-4.1-mini,claude-haiku-4-5-20251001,grok-3-mini',
+    );
+
+    const getModelSpy = jest
+      .spyOn(SERVICE, 'getModel')
+      .mockImplementation((modelId: string) => {
+        if (modelId === 'claude-haiku-4-5-20251001') {
+          return {
+            modelId: 'claude-haiku-4-5-20251001',
+            provider: ModelProvider.ANTHROPIC,
+            model: {} as any,
+          };
+        }
+
+        return undefined;
+      });
+
+    const result = SERVICE.getDefaultSpeedModel();
+
+    expect(result).toBeDefined();
+    expect(result.modelId).toBe('claude-haiku-4-5-20251001');
+    expect(getModelSpy).toHaveBeenCalledWith('gpt-4.1-mini');
+    expect(getModelSpy).toHaveBeenCalledWith('claude-haiku-4-5-20251001');
+  });
+
+  it('should fall back to any available model if none in list are available', () => {
+    MOCK_CONFIG_SERVICE.get.mockReturnValue('model-a,model-b,model-c');
+
+    jest.spyOn(SERVICE, 'getModel').mockReturnValue(undefined);
+    jest.spyOn(SERVICE, 'getAvailableModels').mockReturnValue([
+      {
+        modelId: 'fallback-model',
+        provider: ModelProvider.OPENAI_COMPATIBLE,
+        model: {} as any,
+      },
+    ]);
+
+    const result = SERVICE.getDefaultSpeedModel();
+
+    expect(result).toBeDefined();
+    expect(result.modelId).toBe('fallback-model');
+  });
 });
@@ -1,223 +1,19 @@
-export enum ModelProvider {
-  NONE = 'none',
-  OPENAI = 'openai',
-  ANTHROPIC = 'anthropic',
-  OPENAI_COMPATIBLE = 'open_ai_compatible',
-  XAI = 'xai',
-}
+export {
+  DEFAULT_FAST_MODEL,
+  DEFAULT_SMART_MODEL,
+  ModelProvider,
+  type AIModelConfig,
+  type ModelId,
+  type SupportedFileType,
+} from './ai-models-types.const';

-export const DEFAULT_FAST_MODEL = 'default-fast-model' as const;
-export const DEFAULT_SMART_MODEL = 'default-smart-model' as const;
-
-export type ModelId =
-  | typeof DEFAULT_FAST_MODEL
-  | typeof DEFAULT_SMART_MODEL
-  | 'gpt-4o'
-  | 'gpt-4o-mini'
-  | 'gpt-4-turbo'
-  | 'claude-opus-4-20250514'
-  | 'claude-sonnet-4-20250514'
-  | 'claude-3-5-haiku-20241022'
-  | 'grok-3'
-  | 'grok-3-mini'
-  | 'grok-4'
-  | string; // Allow custom model names
-
-export type SupportedFileType =
-  | 'image/png'
-  | 'image/jpeg'
-  | 'image/gif'
-  | 'image/webp'
-  | 'application/pdf'
-  | 'text/plain'
-  | 'text/html'
-  | 'text/csv'
-  | 'application/json';
-
-export interface AIModelConfig {
-  modelId: ModelId;
-  label: string;
-  description: string;
-  provider: ModelProvider;
-  inputCostPer1kTokensInCents: number;
-  outputCostPer1kTokensInCents: number;
-  contextWindowTokens: number;
-  maxOutputTokens: number;
-  supportedFileTypes?: SupportedFileType[];
-  doesSupportThinking?: boolean;
-  nativeCapabilities?: {
-    webSearch?: boolean;
-    twitterSearch?: boolean;
-  };
-}
+import { type AIModelConfig } from './ai-models-types.const';
+import { ANTHROPIC_MODELS } from './anthropic-models.const';
+import { OPENAI_MODELS } from './openai-models.const';
+import { XAI_MODELS } from './xai-models.const';

 export const AI_MODELS: AIModelConfig[] = [
-  {
-    modelId: 'gpt-4o',
-    label: 'GPT-4o',
-    description:
-      'Most advanced multimodal model with strong reasoning, vision, and coding capabilities',
-    provider: ModelProvider.OPENAI,
-    inputCostPer1kTokensInCents: 0.25,
-    outputCostPer1kTokensInCents: 1.0,
-    contextWindowTokens: 128000,
-    maxOutputTokens: 16384,
-    supportedFileTypes: ['image/png', 'image/jpeg', 'image/gif', 'image/webp'],
-    nativeCapabilities: {
-      webSearch: true,
-    },
-  },
-  {
-    modelId: 'gpt-4o-mini',
-    label: 'GPT-4o Mini',
-    description:
-      'Fast and cost-efficient model for lightweight tasks and high-volume operations',
-    provider: ModelProvider.OPENAI,
-    inputCostPer1kTokensInCents: 0.015,
-    outputCostPer1kTokensInCents: 0.06,
-    contextWindowTokens: 128000,
-    maxOutputTokens: 16384,
-    supportedFileTypes: ['image/png', 'image/jpeg', 'image/gif', 'image/webp'],
-    nativeCapabilities: {
-      webSearch: true,
-    },
-  },
-  {
-    modelId: 'gpt-4-turbo',
-    label: 'GPT-4 Turbo',
-    description:
-      'Previous generation high-performance model with vision capabilities',
-    provider: ModelProvider.OPENAI,
-    inputCostPer1kTokensInCents: 1.0,
-    outputCostPer1kTokensInCents: 3.0,
-    contextWindowTokens: 128000,
-    maxOutputTokens: 4096,
-    supportedFileTypes: ['image/png', 'image/jpeg', 'image/gif', 'image/webp'],
-    nativeCapabilities: {
-      webSearch: false,
-    },
-  },
-  {
-    modelId: 'claude-opus-4-20250514',
-    label: 'Claude Opus 4',
-    description:
-      'Most powerful Claude model with extended thinking for complex reasoning tasks',
-    provider: ModelProvider.ANTHROPIC,
-    inputCostPer1kTokensInCents: 1.5,
-    outputCostPer1kTokensInCents: 7.5,
-    contextWindowTokens: 200000,
-    maxOutputTokens: 8192,
-    supportedFileTypes: [
-      'image/png',
-      'image/jpeg',
-      'image/gif',
-      'image/webp',
-      'application/pdf',
-      'text/plain',
-      'text/html',
-      'text/csv',
-    ],
-    doesSupportThinking: true,
-    nativeCapabilities: {
-      webSearch: true,
-    },
-  },
-  {
-    modelId: 'claude-sonnet-4-20250514',
-    label: 'Claude Sonnet 4',
-    description:
-      'Balanced model with strong performance and extended thinking capabilities',
-    provider: ModelProvider.ANTHROPIC,
-    inputCostPer1kTokensInCents: 0.3,
-    outputCostPer1kTokensInCents: 1.5,
-    contextWindowTokens: 200000,
-    maxOutputTokens: 8192,
-    supportedFileTypes: [
-      'image/png',
-      'image/jpeg',
-      'image/gif',
-      'image/webp',
-      'application/pdf',
-      'text/plain',
-      'text/html',
-      'text/csv',
-    ],
-    doesSupportThinking: true,
-    nativeCapabilities: {
-      webSearch: true,
-    },
-  },
-  {
-    modelId: 'claude-3-5-haiku-20241022',
-    label: 'Claude Haiku 3.5',
-    description:
-      'Fast and efficient model optimized for speed and cost-effectiveness',
-    provider: ModelProvider.ANTHROPIC,
-    inputCostPer1kTokensInCents: 0.08,
-    outputCostPer1kTokensInCents: 0.4,
-    contextWindowTokens: 200000,
-    maxOutputTokens: 8192,
-    supportedFileTypes: [
-      'image/png',
-      'image/jpeg',
-      'image/gif',
-      'image/webp',
-      'application/pdf',
-      'text/plain',
-      'text/html',
-      'text/csv',
-    ],
-    doesSupportThinking: false,
-    nativeCapabilities: {
-      webSearch: true,
-    },
-  },
-  {
-    modelId: 'grok-3',
-    label: 'Grok-3',
-    description:
-      'Advanced model with web and Twitter search, optimized for real-time information',
-    provider: ModelProvider.XAI,
-    inputCostPer1kTokensInCents: 0.3,
-    outputCostPer1kTokensInCents: 1.5,
-    contextWindowTokens: 131072,
-    maxOutputTokens: 8192,
-    supportedFileTypes: ['image/png', 'image/jpeg', 'image/gif', 'image/webp'],
-    nativeCapabilities: {
-      webSearch: true,
-      twitterSearch: true,
-    },
-  },
-  {
-    modelId: 'grok-3-mini',
-    label: 'Grok-3 Mini',
-    description:
-      'Lightweight model with web and Twitter search for fast, cost-effective operations',
-    provider: ModelProvider.XAI,
-    inputCostPer1kTokensInCents: 0.03,
-    outputCostPer1kTokensInCents: 0.05,
-    contextWindowTokens: 131072,
-    maxOutputTokens: 8192,
-    supportedFileTypes: ['image/png', 'image/jpeg', 'image/gif', 'image/webp'],
-    nativeCapabilities: {
-      webSearch: true,
-      twitterSearch: true,
-    },
-  },
-  {
-    modelId: 'grok-4',
-    label: 'Grok-4',
-    description:
-      'Most capable Grok model with enhanced reasoning, web and Twitter search',
-    provider: ModelProvider.XAI,
-    inputCostPer1kTokensInCents: 0.5,
-    outputCostPer1kTokensInCents: 2.5,
-    contextWindowTokens: 131072,
-    maxOutputTokens: 8192,
-    supportedFileTypes: ['image/png', 'image/jpeg', 'image/gif', 'image/webp'],
-    nativeCapabilities: {
-      webSearch: true,
-      twitterSearch: true,
-    },
-  },
+  ...OPENAI_MODELS,
+  ...ANTHROPIC_MODELS,
+  ...XAI_MODELS,
 ];
@@ -0,0 +1,159 @@
+import { type AIModelConfig, ModelProvider } from './ai-models-types.const';
+
+export const ANTHROPIC_MODELS: AIModelConfig[] = [
+  // Active models
+  {
+    modelId: 'claude-opus-4-5-20251101',
+    label: 'Claude Opus 4.5',
+    description:
+      'Most powerful Claude model excelling in complex reasoning, coding, and agentic tasks',
+    provider: ModelProvider.ANTHROPIC,
+    inputCostPer1kTokensInCents: 0.5,
+    outputCostPer1kTokensInCents: 2.5,
+    contextWindowTokens: 200000,
+    maxOutputTokens: 64000,
+    supportedFileTypes: [
+      'image/png',
+      'image/jpeg',
+      'image/gif',
+      'image/webp',
+      'application/pdf',
+      'text/plain',
+      'text/html',
+      'text/csv',
+    ],
+    doesSupportThinking: true,
+    nativeCapabilities: {
+      webSearch: true,
+    },
+  },
+  {
+    modelId: 'claude-sonnet-4-5-20250929',
+    label: 'Claude Sonnet 4.5',
+    description:
+      'Advanced model for coding tasks and complex agent-based workflows with 1M context',
+    provider: ModelProvider.ANTHROPIC,
+    inputCostPer1kTokensInCents: 0.3,
+    outputCostPer1kTokensInCents: 1.5,
+    contextWindowTokens: 1000000,
+    maxOutputTokens: 64000,
+    supportedFileTypes: [
+      'image/png',
+      'image/jpeg',
+      'image/gif',
+      'image/webp',
+      'application/pdf',
+      'text/plain',
+      'text/html',
+      'text/csv',
+    ],
+    doesSupportThinking: true,
+    nativeCapabilities: {
+      webSearch: true,
+    },
+  },
+  {
+    modelId: 'claude-haiku-4-5-20251001',
+    label: 'Claude Haiku 4.5',
+    description:
+      'Fast and cost-effective model optimized for high-speed processing',
+    provider: ModelProvider.ANTHROPIC,
+    inputCostPer1kTokensInCents: 0.1,
+    outputCostPer1kTokensInCents: 0.5,
+    contextWindowTokens: 200000,
+    maxOutputTokens: 64000,
+    supportedFileTypes: [
+      'image/png',
+      'image/jpeg',
+      'image/gif',
+      'image/webp',
+      'application/pdf',
+      'text/plain',
+      'text/html',
+      'text/csv',
+    ],
+    doesSupportThinking: false,
+    nativeCapabilities: {
+      webSearch: true,
+    },
+  },
+  {
+    modelId: 'claude-3-5-haiku-20241022',
+    label: 'Claude Haiku 3.5',
+    description:
+      'Fast and efficient model optimized for speed and cost-effectiveness',
+    provider: ModelProvider.ANTHROPIC,
+    inputCostPer1kTokensInCents: 0.08,
+    outputCostPer1kTokensInCents: 0.4,
+    contextWindowTokens: 200000,
+    maxOutputTokens: 8192,
+    supportedFileTypes: [
+      'image/png',
+      'image/jpeg',
+      'image/gif',
+      'image/webp',
+      'application/pdf',
+      'text/plain',
+      'text/html',
+      'text/csv',
+    ],
+    doesSupportThinking: false,
+    nativeCapabilities: {
+      webSearch: true,
+    },
+  },
+
+  // Deprecated models - kept for backward compatibility with existing agents
+  {
+    modelId: 'claude-opus-4-20250514',
+    label: 'Claude Opus 4',
+    description:
+      'Most powerful Claude model with extended thinking for complex reasoning tasks',
+    provider: ModelProvider.ANTHROPIC,
+    inputCostPer1kTokensInCents: 1.5,
+    outputCostPer1kTokensInCents: 7.5,
+    contextWindowTokens: 200000,
+    maxOutputTokens: 8192,
+    supportedFileTypes: [
+      'image/png',
+      'image/jpeg',
+      'image/gif',
+      'image/webp',
+      'application/pdf',
+      'text/plain',
+      'text/html',
+      'text/csv',
+    ],
+    doesSupportThinking: true,
+    nativeCapabilities: {
+      webSearch: true,
+    },
+    deprecated: true,
+  },
+  {
+    modelId: 'claude-sonnet-4-20250514',
+    label: 'Claude Sonnet 4',
+    description:
+      'Balanced model with strong performance and extended thinking capabilities',
+    provider: ModelProvider.ANTHROPIC,
+    inputCostPer1kTokensInCents: 0.3,
+    outputCostPer1kTokensInCents: 1.5,
+    contextWindowTokens: 200000,
+    maxOutputTokens: 8192,
+    supportedFileTypes: [
+      'image/png',
+      'image/jpeg',
+      'image/gif',
+      'image/webp',
+      'application/pdf',
+      'text/plain',
+      'text/html',
+      'text/csv',
+    ],
+    doesSupportThinking: true,
+    nativeCapabilities: {
+      webSearch: true,
+    },
+    deprecated: true,
+  },
+];
@@ -0,0 +1,117 @@
+import { type AIModelConfig, ModelProvider } from './ai-models-types.const';
+
+export const OPENAI_MODELS: AIModelConfig[] = [
+  // Active models
+  {
+    modelId: 'gpt-4.1',
+    label: 'GPT-4.1',
+    description:
+      'Advanced model excelling in coding, instruction following, and long-context comprehension',
+    provider: ModelProvider.OPENAI,
+    inputCostPer1kTokensInCents: 0.2,
+    outputCostPer1kTokensInCents: 0.8,
+    contextWindowTokens: 1047576,
+    maxOutputTokens: 32768,
+    supportedFileTypes: ['image/png', 'image/jpeg', 'image/gif', 'image/webp'],
+    nativeCapabilities: {
+      webSearch: true,
+    },
+  },
+  {
+    modelId: 'gpt-4.1-mini',
+    label: 'GPT-4.1 Mini',
+    description:
+      'Fast and cost-efficient version of GPT-4.1 optimized for low latency',
+    provider: ModelProvider.OPENAI,
+    inputCostPer1kTokensInCents: 0.04,
+    outputCostPer1kTokensInCents: 0.16,
+    contextWindowTokens: 1047576,
+    maxOutputTokens: 32768,
+    supportedFileTypes: ['image/png', 'image/jpeg', 'image/gif', 'image/webp'],
+    nativeCapabilities: {
+      webSearch: true,
+    },
+  },
+  {
+    modelId: 'o3',
+    label: 'o3',
+    description:
+      'Powerful reasoning model excelling in complex queries, coding, math, and science',
+    provider: ModelProvider.OPENAI,
+    inputCostPer1kTokensInCents: 0.2,
+    outputCostPer1kTokensInCents: 0.8,
+    contextWindowTokens: 200000,
+    maxOutputTokens: 100000,
+    supportedFileTypes: ['image/png', 'image/jpeg', 'image/gif', 'image/webp'],
+    doesSupportThinking: true,
+    nativeCapabilities: {
+      webSearch: true,
+    },
+  },
+  {
+    modelId: 'o4-mini',
+    label: 'o4-mini',
+    description:
+      'Cost-effective reasoning model excelling in math, coding, and visual tasks',
+    provider: ModelProvider.OPENAI,
+    inputCostPer1kTokensInCents: 0.11,
+    outputCostPer1kTokensInCents: 0.44,
+    contextWindowTokens: 200000,
+    maxOutputTokens: 100000,
+    supportedFileTypes: ['image/png', 'image/jpeg', 'image/gif', 'image/webp'],
+    doesSupportThinking: true,
+    nativeCapabilities: {
+      webSearch: true,
+    },
+  },
+
+  // Deprecated models - kept for backward compatibility with existing agents
+  {
+    modelId: 'gpt-4o',
+    label: 'GPT-4o',
+    description:
+      'Most advanced multimodal model with strong reasoning, vision, and coding capabilities',
+    provider: ModelProvider.OPENAI,
+    inputCostPer1kTokensInCents: 0.25,
+    outputCostPer1kTokensInCents: 1.0,
+    contextWindowTokens: 128000,
+    maxOutputTokens: 16384,
+    supportedFileTypes: ['image/png', 'image/jpeg', 'image/gif', 'image/webp'],
+    nativeCapabilities: {
+      webSearch: true,
+    },
+    deprecated: true,
+  },
+  {
+    modelId: 'gpt-4o-mini',
+    label: 'GPT-4o Mini',
+    description:
+      'Fast and cost-efficient model for lightweight tasks and high-volume operations',
+    provider: ModelProvider.OPENAI,
+    inputCostPer1kTokensInCents: 0.015,
+    outputCostPer1kTokensInCents: 0.06,
+    contextWindowTokens: 128000,
+    maxOutputTokens: 16384,
+    supportedFileTypes: ['image/png', 'image/jpeg', 'image/gif', 'image/webp'],
+    nativeCapabilities: {
+      webSearch: true,
+    },
+    deprecated: true,
+  },
+  {
+    modelId: 'gpt-4-turbo',
+    label: 'GPT-4 Turbo',
+    description:
+      'Previous generation high-performance model with vision capabilities',
+    provider: ModelProvider.OPENAI,
+    inputCostPer1kTokensInCents: 1.0,
+    outputCostPer1kTokensInCents: 3.0,
+    contextWindowTokens: 128000,
+    maxOutputTokens: 4096,
+    supportedFileTypes: ['image/png', 'image/jpeg', 'image/gif', 'image/webp'],
+    nativeCapabilities: {
+      webSearch: false,
+    },
+    deprecated: true,
+  },
+];
@@ -0,0 +1,70 @@
+import { type AIModelConfig, ModelProvider } from './ai-models-types.const';
+
+export const XAI_MODELS: AIModelConfig[] = [
+  // Active models
+  {
+    modelId: 'grok-4-1-fast-reasoning',
+    label: 'Grok 4.1 Fast',
+    description:
+      'Next-generation tool-calling agent with 2M context for advanced agentic workflows',
+    provider: ModelProvider.XAI,
+    inputCostPer1kTokensInCents: 0.02,
+    outputCostPer1kTokensInCents: 0.05,
+    contextWindowTokens: 2000000,
+    maxOutputTokens: 8192,
+    supportedFileTypes: ['image/png', 'image/jpeg', 'image/gif', 'image/webp'],
+    doesSupportThinking: true,
+    nativeCapabilities: {
+      webSearch: true,
+      twitterSearch: true,
+    },
+  },
+  {
+    modelId: 'grok-4',
+    label: 'Grok-4',
+    description:
+      'Most capable Grok model with enhanced reasoning, web and Twitter search',
+    provider: ModelProvider.XAI,
+    inputCostPer1kTokensInCents: 0.3,
+    outputCostPer1kTokensInCents: 1.5,
+    contextWindowTokens: 256000,
+    maxOutputTokens: 8192,
+    supportedFileTypes: ['image/png', 'image/jpeg', 'image/gif', 'image/webp'],
+    nativeCapabilities: {
+      webSearch: true,
+      twitterSearch: true,
+    },
+  },
+  {
+    modelId: 'grok-3',
+    label: 'Grok-3',
+    description:
+      'Advanced model with web and Twitter search, optimized for real-time information',
+    provider: ModelProvider.XAI,
+    inputCostPer1kTokensInCents: 0.3,
+    outputCostPer1kTokensInCents: 1.5,
+    contextWindowTokens: 131072,
+    maxOutputTokens: 8192,
+    supportedFileTypes: ['image/png', 'image/jpeg', 'image/gif', 'image/webp'],
+    nativeCapabilities: {
+      webSearch: true,
+      twitterSearch: true,
+    },
+  },
+  {
+    modelId: 'grok-3-mini',
+    label: 'Grok-3 Mini',
+    description:
+      'Lightweight model with web and Twitter search for fast, cost-effective operations',
+    provider: ModelProvider.XAI,
+    inputCostPer1kTokensInCents: 0.03,
+    outputCostPer1kTokensInCents: 0.05,
+    contextWindowTokens: 131072,
+    maxOutputTokens: 8192,
+    supportedFileTypes: ['image/png', 'image/jpeg', 'image/gif', 'image/webp'],
+    nativeCapabilities: {
+      webSearch: true,
+      twitterSearch: true,
+    },
+  },
+];
@@ -13,6 +13,9 @@ import {
  ModelProvider,
  type AIModelConfig,
 } from 'src/engine/metadata-modules/ai/ai-models/constants/ai-models.const';
+import { ANTHROPIC_MODELS } from 'src/engine/metadata-modules/ai/ai-models/constants/anthropic-models.const';
+import { OPENAI_MODELS } from 'src/engine/metadata-modules/ai/ai-models/constants/openai-models.const';
+import { XAI_MODELS } from 'src/engine/metadata-modules/ai/ai-models/constants/xai-models.const';

 export interface RegisteredAIModel {
  modelId: string;
@@ -66,25 +69,18 @@ export class AiModelRegistryService {
  }

  private registerOpenAIModels(): void {
-    const openaiModels = AI_MODELS.filter(
-      (model) => model.provider === ModelProvider.OPENAI,
-    );
-
-    openaiModels.forEach((modelConfig) => {
+    OPENAI_MODELS.forEach((modelConfig) => {
      this.modelRegistry.set(modelConfig.modelId, {
        modelId: modelConfig.modelId,
        provider: ModelProvider.OPENAI,
        model: openai(modelConfig.modelId),
+        doesSupportThinking: modelConfig.doesSupportThinking,
      });
    });
  }

  private registerAnthropicModels(): void {
-    const anthropicModels = AI_MODELS.filter(
-      (model) => model.provider === ModelProvider.ANTHROPIC,
-    );
-
-    anthropicModels.forEach((modelConfig) => {
+    ANTHROPIC_MODELS.forEach((modelConfig) => {
      this.modelRegistry.set(modelConfig.modelId, {
        modelId: modelConfig.modelId,
        provider: ModelProvider.ANTHROPIC,
@@ -95,15 +91,12 @@ export class AiModelRegistryService {
  }

  private registerXaiModels(): void {
-    const xaiModels = AI_MODELS.filter(
-      (model) => model.provider === ModelProvider.XAI,
-    );
-
-    xaiModels.forEach((modelConfig) => {
+    XAI_MODELS.forEach((modelConfig) => {
      this.modelRegistry.set(modelConfig.modelId, {
        modelId: modelConfig.modelId,
        provider: ModelProvider.XAI,
        model: xai(modelConfig.modelId),
+        doesSupportThinking: modelConfig.doesSupportThinking,
      });
    });
  }
@@ -140,11 +133,30 @@ export class AiModelRegistryService {
    return Array.from(this.modelRegistry.values());
  }

+  private getFirstAvailableModelFromList(
+    modelIdList: string,
+  ): RegisteredAIModel | undefined {
+    const modelIds = modelIdList
+      .split(',')
+      .map((id) => id.trim())
+      .filter((id) => id.length > 0);
+
+    for (const modelId of modelIds) {
+      const model = this.getModel(modelId);
+
+      if (model) {
+        return model;
+      }
+    }
+
+    return undefined;
+  }
+
  getDefaultSpeedModel(): RegisteredAIModel {
-    const defaultModelId = this.twentyConfigService.get(
+    const defaultModelIds = this.twentyConfigService.get(
      'DEFAULT_AI_SPEED_MODEL_ID',
    );
-    let model = this.getModel(defaultModelId);
+    let model = this.getFirstAvailableModelFromList(defaultModelIds);

    if (!model) {
      const availableModels = this.getAvailableModels();
@@ -156,10 +168,10 @@ export class AiModelRegistryService {
  }

  getDefaultPerformanceModel(): RegisteredAIModel {
-    const defaultModelId = this.twentyConfigService.get(
+    const defaultModelIds = this.twentyConfigService.get(
      'DEFAULT_AI_PERFORMANCE_MODEL_ID',
    );
-    let model = this.getModel(defaultModelId);
+    let model = this.getFirstAvailableModelFromList(defaultModelIds);

    if (!model) {
      const availableModels = this.getAvailableModels();