@tyvm/knowhow 0.0.104 → 0.0.106

package/src/clients/pricing/google.ts

@@ -1,236 +1,275 @@
-import { Models, EmbeddingModels } from "../../types";
-
 /**
- * Gemini model pricing per 1M tokens (USD).
- *
- * Many models have modality-dependent pricing:
- * - `input`: text / image / video input rate
- * - `input_audio`: audio input rate (if different from text)
- * - `output`: text output rate
- * - `output_audio`: audio output rate (if different from text)
- * - `context_caching`: context cache read rate (text/image/video)
- * - `context_caching_audio`: context cache read rate for audio tokens
- * - `context_caching_gt_200k`: rate for prompts > 200k tokens
- * - `context_caching_storage`: storage cost per 1M tokens per hour
- * - `image_generation`: per-image cost for image output models
- * - `video_generation`: per-second cost for video generation models
+ * Google model IDs, pricing, and catalog.
+ * Single source of truth for all Google/Gemini models.
  */
-export interface GeminiModelPricing {
-  input?: number;
+import { completions, embeddings, images, videos, audios, liveApi, ModelCatalogEntry, ModelPricing } from "./types";
+
+// ─── Model IDs ────────────────────────────────────────────────────────────────
+
+export const GoogleModels = {
+  // Gemini 3.x
+  Gemini_31_Pro_Preview: "gemini-3.1-pro-preview",
+  Gemini_31_Flash_Image_Preview: "gemini-3.1-flash-image-preview",
+  Gemini_31_Flash_Lite_Preview: "gemini-3.1-flash-lite-preview",
+  Gemini_31_Flash_Live_Preview: "gemini-3.1-flash-live-preview",
+  Gemini_3_Flash_Preview: "gemini-3-flash-preview",
+  Gemini_3_Pro_Image_Preview: "gemini-3-pro-image-preview",
+  // Gemini 2.5
+  Gemini_25_Pro: "gemini-2.5-pro",
+  Gemini_25_Flash: "gemini-2.5-flash",
+  Gemini_25_Flash_Lite: "gemini-2.5-flash-lite",
+  Gemini_25_Flash_Preview: "gemini-2.5-flash-preview-05-20",
+  Gemini_25_Flash_Preview_0417: "gemini-2.5-flash-preview-04-17",
+  Gemini_25_Pro_Preview: "gemini-2.5-pro-preview-05-06",
+  Gemini_25_Flash_Image: "gemini-2.5-flash-image",
+  Gemini_25_Flash_Image_Preview: "gemini-2.5-flash-image-preview",
+  Gemini_25_Flash_Live: "gemini-2.5-flash-live-preview",
+  Gemini_25_Flash_Native_Audio: "gemini-2.5-flash-native-audio-preview-12-2025",
+  Gemini_25_Pro_TTS: "gemini-2.5-pro-preview-tts",
+  // Gemini 2.0
+  Gemini_20_Flash: "gemini-2.0-flash",
+  Gemini_20_Flash_Preview_Image_Generation: "gemini-2.0-flash-exp-image-generation",
+  Gemini_20_Flash_Live: "gemini-2.0-flash-live-001",
+  // Gemini 1.5 (legacy)
+  Gemini_15_Flash: "gemini-1.5-flash",
+  Gemini_15_Flash_8B: "gemini-1.5-flash-8b",
+  Gemini_15_Pro: "gemini-1.5-pro",
+  // Media generation
+  Imagen_3: "imagen-4.0-generate-001",
+  Imagen_4_Fast: "imagen-4.0-fast-generate-001",
+  Imagen_4_Ultra: "imagen-4.0-ultra-generate-001",
+  Veo_2: "veo-2.0-generate-001",
+  Veo_3: "veo-3.0-generate-001",
+  Veo_3_Fast: "veo-3.0-fast-generate-001",
+  Veo_3_1: "veo-3.1-generate-preview",
+  Veo_3_1_Fast: "veo-3.1-fast-generate-preview",
+  // Audio / TTS
+  Gemini_25_Flash_TTS: "gemini-2.5-flash-preview-tts",
+  Gemini_20_Flash_TTS: "gemini-2.0-flash-preview-tts",
+} as const;
+
+// ─── Gemma models (free via Google AI / NVIDIA NIM) ───────────────────────────
+export const GemmaModels = {
+  Gemma_3_27B_It: "gemma-3-27b-it",
+  Gemma_3_12B_It: "gemma-3-12b-it",
+  Gemma_3_4B_It: "gemma-3-4b-it",
+  Gemma_3n_E2B_It: "gemma-3n-e2b-it",
+  Gemma_3n_E4B_It: "gemma-3n-e4b-it",
+  Gemma_4_26B_A4B_It: "gemma-4-26b-a4b-it",
+  Gemma_4_31B_It: "gemma-4-31b-it",
+} as const;
+
+export const GoogleEmbeddingModels: Record<string, string> = {
+  Gemini_Embedding: "gemini-embedding-exp",
+  Gemini_Embedding_001: "gemini-embedding-001",
+};
+
+// ─── Modality arrays ──────────────────────────────────────────────────────────
+
+export const GoogleTextModels: string[] = [
+  GoogleModels.Gemini_31_Pro_Preview,
+  GoogleModels.Gemini_31_Flash_Lite_Preview,
+  GoogleModels.Gemini_25_Pro,
+  GoogleModels.Gemini_25_Flash,
+  GoogleModels.Gemini_25_Flash_Lite,
+  GoogleModels.Gemini_20_Flash,
+];
+
+// Models that support thinkingLevel (Gemini 3.x series)
+export const GoogleThinkingLevelModels: string[] = [
+  GoogleModels.Gemini_31_Pro_Preview,
+  GoogleModels.Gemini_31_Flash_Lite_Preview,
+  GoogleModels.Gemini_3_Flash_Preview,
+];
+
+// Models that support thinkingBudget (Gemini 2.5 series)
+export const GoogleThinkingBudgetModels: string[] = [
+  GoogleModels.Gemini_25_Pro,
+  GoogleModels.Gemini_25_Flash,
+  GoogleModels.Gemini_25_Flash_Lite,
+  GoogleModels.Gemini_25_Flash_Preview,
+  GoogleModels.Gemini_25_Flash_Preview_0417,
+  GoogleModels.Gemini_25_Pro_Preview,
+];
+
+// Live API only — not compatible with generateContent (text completions)
+export const GoogleLiveApiModels: string[] = [
+  GoogleModels.Gemini_31_Flash_Live_Preview,
+  GoogleModels.Gemini_25_Flash_Live,
+  GoogleModels.Gemini_25_Flash_Native_Audio,
+  GoogleModels.Gemini_20_Flash_Live,
+];
+
+// Limited availability — exist in catalog but return empty or restricted responses
+export const GoogleLimitedModels: string[] = [
+  GoogleModels.Gemini_3_Flash_Preview,
+];
+
+// All deprecated/legacy models — metadata is embedded in GeminiPricing entries
+const GoogleAllDeprecatedModels: string[] = [
+  // Deprecated previews (shutdownDate + replacedBy in GeminiPricing)
+  GoogleModels.Gemini_25_Flash_Preview,
+  GoogleModels.Gemini_25_Flash_Preview_0417,
+  GoogleModels.Gemini_25_Pro_Preview,
+  // Legacy (Gemini 1.5 — returning 404)
+  GoogleModels.Gemini_15_Flash,
+  GoogleModels.Gemini_15_Flash_8B,
+  GoogleModels.Gemini_15_Pro,
+];
+
+export const GemmaTextModels: string[] = [
+  GemmaModels.Gemma_3_27B_It,
+  GemmaModels.Gemma_3_12B_It,
+  GemmaModels.Gemma_3_4B_It,
+  GemmaModels.Gemma_3n_E2B_It,
+  GemmaModels.Gemma_3n_E4B_It,
+  GemmaModels.Gemma_4_26B_A4B_It,
+  GemmaModels.Gemma_4_31B_It,
+];
+
+export const GoogleImageModels: string[] = [
+  GoogleModels.Gemini_31_Flash_Image_Preview,
+  GoogleModels.Gemini_3_Pro_Image_Preview,
+  GoogleModels.Gemini_25_Flash_Image,
+  GoogleModels.Gemini_25_Flash_Image_Preview,
+  GoogleModels.Gemini_20_Flash_Preview_Image_Generation,
+  GoogleModels.Imagen_3,
+  GoogleModels.Imagen_4_Fast,
+  GoogleModels.Imagen_4_Ultra,
+];
+
+export const GoogleVideoModels: string[] = [
+  GoogleModels.Veo_2,
+  GoogleModels.Veo_3,
+  GoogleModels.Veo_3_Fast,
+  GoogleModels.Veo_3_1,
+  GoogleModels.Veo_3_1_Fast,
+];
+
+export const GoogleTTSModels: string[] = [
+  GoogleModels.Gemini_25_Flash_TTS,
+  GoogleModels.Gemini_25_Pro_TTS,
+  GoogleModels.Gemini_20_Flash_TTS,
+];
+
+export const GoogleEmbeddingModelsList: string[] = [
+  GoogleEmbeddingModels.Gemini_Embedding,
+  GoogleEmbeddingModels.Gemini_Embedding_001,
+];
+
+// ─── Pricing ──────────────────────────────────────────────────────────────────
+
+export interface GeminiModelPricing extends ModelPricing {
   input_audio?: number;
   input_gt_200k?: number;
-  output?: number;
-  output_audio?: number;
   output_gt_200k?: number;
   context_caching?: number;
   context_caching_audio?: number;
   context_caching_gt_200k?: number;
-  context_caching_storage?: number; // per 1M tokens per hour
-  image_generation?: number;
-  video_generation?: number;
+  context_caching_storage?: number;
   thinking_output?: number;
 }
 
 export const GeminiPricing: Record<string, GeminiModelPricing> = {
-  // ── Gemini 3.x ────────────────────────────────────────────────────────────
-  [Models.google.Gemini_31_Pro_Preview]: {
-    input: 2,
-    input_gt_200k: 4,
-    output: 12,
-    output_gt_200k: 18,
-    context_caching: 0.2,
-    context_caching_gt_200k: 0.4,
-    context_caching_storage: 4.5,
-  },
-  // gemini-3.1-flash-image-preview: text/image input, text+image output
-  [Models.google.Gemini_31_Flash_Image_Preview]: {
-    input: 0.5,
-    output: 3,
-    image_generation: 0.045, // per 0.5K image ($60/1M tokens × 747 tokens)
+  [GoogleModels.Gemini_31_Pro_Preview]: {
+    input: 2, input_gt_200k: 4, output: 12, output_gt_200k: 18,
+    context_caching: 0.2, context_caching_gt_200k: 0.4, context_caching_storage: 4.5,
   },
-  // gemini-3.1-flash-lite-preview: audio input costs 2× text
-  [Models.google.Gemini_31_Flash_Lite_Preview]: {
-    input: 0.25,
-    input_audio: 0.5,
-    output: 1.5,
-    context_caching: 0.025,
-    context_caching_audio: 0.05,
-    context_caching_storage: 1.0,
+  [GoogleModels.Gemini_31_Flash_Image_Preview]: {
+    input: 0.25, output: 60.0, image_generation: 0.045, image_generation_per_1m_tokens: 60.0,
   },
-  // gemini-3.1-flash-live-preview: Live API, per-modality rates
-  // Input: $0.75 text, $3.00 audio/video, $1.00 image
-  // Output: $4.50 text, $12.00 audio
-  [Models.google.Gemini_31_Flash_Live_Preview]: {
-    input: 0.75,        // text
-    input_audio: 3.0,   // audio / video
-    output: 4.5,        // text
-    output_audio: 12.0, // audio
+  [GoogleModels.Gemini_31_Flash_Lite_Preview]: {
+    input: 0.25, input_audio: 0.5, output: 1.5,
+    context_caching: 0.025, context_caching_audio: 0.05, context_caching_storage: 1.0,
   },
-  // gemini-3-flash-preview: audio input costs 2× text
-  [Models.google.Gemini_3_Flash_Preview]: {
-    input: 0.5,
-    input_audio: 1.0,
-    output: 3.0,
-    context_caching: 0.05,
-    context_caching_audio: 0.10,
-    context_caching_storage: 1.0,
+  [GoogleModels.Gemini_31_Flash_Live_Preview]: {
+    input: 0.75, input_audio: 3.0, output: 4.5, output_audio: 12.0,
   },
-  // gemini-3-pro-image-preview: text+image input, text+image output
-  [Models.google.Gemini_3_Pro_Image_Preview]: {
-    input: 2,
-    output: 12,
-    image_generation: 0.134, // per 1K/2K image ($120/1M tokens × 1120 tokens)
+  [GoogleModels.Gemini_3_Flash_Preview]: {
+    input: 0.5, input_audio: 1.0, output: 3.0,
+    context_caching: 0.05, context_caching_audio: 0.10, context_caching_storage: 1.0,
+    limitedAvailability: true,
   },
-
-  // ── Gemini 2.5 ────────────────────────────────────────────────────────────
-  [Models.google.Gemini_25_Pro]: {
-    input: 1.25,
-    input_gt_200k: 2.5,
-    output: 10.0,
-    output_gt_200k: 15.0,
-    context_caching: 0.125,
-    context_caching_gt_200k: 0.25,
-    context_caching_storage: 4.5,
-  },
-  // gemini-2.5-flash: audio input costs 3.3× text; audio caching 3.3× text
-  [Models.google.Gemini_25_Flash]: {
-    input: 0.3,
-    input_audio: 1.0,
-    output: 2.5,
-    context_caching: 0.03,
-    context_caching_audio: 0.1,
-    context_caching_storage: 1.0,
+  [GoogleModels.Gemini_3_Pro_Image_Preview]: {
+    input: 2, output: 12, image_generation: 0.134,
   },
-  // gemini-2.5-flash-lite: audio input costs 3× text; audio caching 3× text
-  [Models.google.Gemini_25_Flash_Lite]: {
-    input: 0.1,
-    input_audio: 0.3,
-    output: 0.4,
-    context_caching: 0.01,
-    context_caching_audio: 0.03,
-    context_caching_storage: 1.0,
+  [GoogleModels.Gemini_25_Pro]: {
+    input: 1.25, input_gt_200k: 2.5, output: 10.0, output_gt_200k: 15.0,
+    context_caching: 0.125, context_caching_gt_200k: 0.25, context_caching_storage: 4.5,
   },
-  // gemini-2.5-flash-preview-05-20: same modality splits as 2.5 Flash
-  [Models.google.Gemini_25_Flash_Preview]: {
-    input: 0.3,
-    input_audio: 1.0,
-    output: 2.5,
-    context_caching: 0.03,
-    context_caching_audio: 0.1,
-    context_caching_storage: 1.0,
+  [GoogleModels.Gemini_25_Flash]: {
+    input: 0.3, input_audio: 1.0, output: 2.5,
+    context_caching: 0.03, context_caching_audio: 0.1, context_caching_storage: 1.0,
   },
-  [Models.google.Gemini_25_Pro_Preview]: {
-    input: 1.25,
-    input_gt_200k: 2.5,
-    output: 10.0,
-    output_gt_200k: 15.0,
-    context_caching: 0.125,
-    context_caching_gt_200k: 0.25,
-    context_caching_storage: 4.5,
+  [GoogleModels.Gemini_25_Flash_Lite]: {
+    input: 0.1, input_audio: 0.3, output: 0.4,
+    context_caching: 0.01, context_caching_audio: 0.03, context_caching_storage: 1.0,
   },
-  // gemini-2.5-flash-image: image output model, per-image pricing
-  [Models.google.Gemini_25_Flash_Image]: {
-    input: 0.3,
-    output: 0.039, // per image ($30/1M tokens × 1290 tokens)
+  [GoogleModels.Gemini_25_Flash_Preview_0417]: { input: 0.15, output: 0.60, context_caching: 0.0375, deprecated: true, deprecationDate: "2025-11-18", replacedBy: GoogleModels.Gemini_3_Flash_Preview },
+  [GoogleModels.Gemini_25_Flash_Preview]: {
+    input: 0.15, input_audio: 0.5, output: 0.60, context_caching: 0.0375, deprecated: true, deprecationDate: "2025-11-18", replacedBy: GoogleModels.Gemini_3_Flash_Preview,
   },
-  // gemini-2.5-flash-live / gemini-2.5-flash-native-audio-preview-12-2025:
-  // Live API models — audio I/O costs dramatically more than text
-  // Input: $0.50 text, $3.00 audio/video
-  // Output: $2.00 text, $12.00 audio
-  [Models.google.Gemini_25_Flash_Live]: {
-    input: 0.5,
-    input_audio: 3.0,
-    output: 2.0,
-    output_audio: 12.0,
+  [GoogleModels.Gemini_25_Pro_Preview]: {
+    input: 1.25, input_gt_200k: 2.5, output: 10.0, output_gt_200k: 15.0,
+    context_caching: 0.125, context_caching_gt_200k: 0.25, context_caching_storage: 4.5,
+    deprecated: true, deprecationDate: "2025-12-02", replacedBy: GoogleModels.Gemini_31_Pro_Preview,
   },
-  [Models.google.Gemini_25_Flash_Native_Audio]: {
-    input: 0.5,
-    input_audio: 3.0,
-    output: 2.0,
-    output_audio: 12.0,
+  [GoogleModels.Gemini_25_Flash_Image]: {
+    input: 0.3, output: 30.0, image_generation: 0.039, image_generation_per_1m_tokens: 30.0,
   },
-  // TTS models: text-only input, audio-only output
-  [Models.google.Gemini_25_Flash_TTS]: {
-    input: 0.5,
-    output_audio: 10.0,
+  [GoogleModels.Gemini_25_Flash_Image_Preview]: {
+    input: 0.3, output: 30.0, image_generation: 0.039, image_generation_per_1m_tokens: 30.0,
   },
-  [Models.google.Gemini_25_Pro_TTS]: {
-    input: 1.0,
-    output_audio: 20.0,
-  },
-
-  // ── Gemini 2.0 (deprecated) ───────────────────────────────────────────────
-  // gemini-2.0-flash: audio input costs 7× text; audio caching costs 7× text
-  [Models.google.Gemini_20_Flash]: {
-    input: 0.1,
-    input_audio: 0.7,
-    output: 0.4,
-    context_caching: 0.025,
-    context_caching_audio: 0.175,
-    context_caching_storage: 1.0,
-  },
-  [Models.google.Gemini_20_Flash_Preview_Image_Generation]: {
-    input: 0.1,
-    output: 0.4,
-    image_generation: 0.039,
-  },
-
-  // ── Gemini 1.5 (legacy) ───────────────────────────────────────────────────
-  [Models.google.Gemini_15_Flash]: {
-    input: 0.075,
-    output: 0.3,
-    context_caching: 0.01875,
-  },
-  [Models.google.Gemini_15_Flash_8B]: {
-    input: 0.0375,
-    output: 0.15,
-    context_caching: 0.01,
-  },
-  [Models.google.Gemini_15_Pro]: {
-    input: 1.25,
-    output: 5.0,
-    context_caching: 0.3125,
-  },
-
-  // ── Image generation ──────────────────────────────────────────────────────
-  [Models.google.Imagen_3]: {
-    image_generation: 0.04, // $0.04/image (Imagen 4 Standard)
-  },
-  [Models.google.Imagen_4_Fast]: {
-    image_generation: 0.02,
-  },
-  [Models.google.Imagen_4_Ultra]: {
-    image_generation: 0.06,
-  },
-
-  // ── Video generation ──────────────────────────────────────────────────────
-  [Models.google.Veo_2]: {
-    video_generation: 0.35, // per second
-  },
-  [Models.google.Veo_3]: {
-    video_generation: 0.4,
-  },
-  [Models.google.Veo_3_Fast]: {
-    video_generation: 0.1, // 720p base rate; 1080p=$0.12, 4k=$0.30 not expressible here
-  },
-  [Models.google.Veo_3_1]: {
-    video_generation: 0.4, // 720p/1080p; 4k=$0.60 not expressible as single scalar
-  },
-  [Models.google.Veo_3_1_Fast]: {
-    video_generation: 0.1, // 720p base; 1080p=$0.12, 4k=$0.30
+  [GoogleModels.Gemini_25_Flash_Live]: { input: 0.5, input_audio: 3.0, output: 2.0, output_audio: 12.0 },
+  [GoogleModels.Gemini_25_Flash_Native_Audio]: { input: 0.5, input_audio: 3.0, output: 2.0, output_audio: 12.0 },
+  [GoogleModels.Gemini_25_Flash_TTS]: { input: 0.5, output_audio: 10.0, output: 10.0 },
+  [GoogleModels.Gemini_25_Pro_TTS]: { input: 1.0, output_audio: 20.0, output: 20.0 },
+  [GoogleModels.Gemini_20_Flash_TTS]: { input: 0.1, output_audio: 4.0, output: 4.0 },
+  [GoogleModels.Gemini_20_Flash]: {
+    input: 0.1, input_audio: 0.7, output: 0.4,
+    context_caching: 0.025, context_caching_audio: 0.175, context_caching_storage: 1.0,
   },
+  [GoogleModels.Gemini_20_Flash_Preview_Image_Generation]: { input: 0.1, output: 0.4, image_generation: 0.039 },
+  [GoogleModels.Gemini_20_Flash_Live]: { input: 0.1, output: 0.4 },
+  [GoogleModels.Gemini_15_Flash]:   { input: 0.075, output: 0.3,  context_caching: 0.01875, deprecated: true },
+  [GoogleModels.Gemini_15_Flash_8B]:{ input: 0.0375, output: 0.15, context_caching: 0.01,   deprecated: true },
+  [GoogleModels.Gemini_15_Pro]:     { input: 1.25, output: 5.0,   context_caching: 0.3125,  deprecated: true },
+  [GoogleModels.Imagen_3]: { image_generation: 0.04 },
+  [GoogleModels.Imagen_4_Fast]: { image_generation: 0.02 },
+  [GoogleModels.Imagen_4_Ultra]: { image_generation: 0.06 },
+  [GoogleModels.Veo_2]: { video_generation: 0.35 },
+  [GoogleModels.Veo_3]: { video_generation: 0.4 },
+  [GoogleModels.Veo_3_Fast]: { video_generation: 0.1 },
+  [GoogleModels.Veo_3_1]: { video_generation: 0.4 },
+  [GoogleModels.Veo_3_1_Fast]: { video_generation: 0.1 },
+  [GoogleEmbeddingModels.Gemini_Embedding]: { input: 0 },
+  [GoogleEmbeddingModels.Gemini_Embedding_001]: { input: 0.15 },
+};
 
-  // ── Embeddings ────────────────────────────────────────────────────────────
-  [EmbeddingModels.google.Gemini_Embedding]: {
-    input: 0, // Free of charge
-  },
-  [EmbeddingModels.google.Gemini_Embedding_001]: {
-    input: 0.15,
-  },
+// ─── Gemma pricing (free via Google AI) ──────────────────────────────────────
+export const GemmaPricing: Record<string, GeminiModelPricing> = {
+  [GemmaModels.Gemma_3_27B_It]: { input: 0.0, output: 0.0 },
+  [GemmaModels.Gemma_3_12B_It]: { input: 0.0, output: 0.0 },
+  [GemmaModels.Gemma_3_4B_It]: { input: 0.0, output: 0.0 },
+  [GemmaModels.Gemma_3n_E2B_It]: { input: 0.0, output: 0.0 },
+  [GemmaModels.Gemma_3n_E4B_It]: { input: 0.0, output: 0.0 },
+  [GemmaModels.Gemma_4_26B_A4B_It]: { input: 0.0, output: 0.0 },
+  [GemmaModels.Gemma_4_31B_It]: { input: 0.0, output: 0.0 },
 };
 
-/**
- * @deprecated Use GeminiPricing instead.
- * Kept as alias for backwards compatibility.
- */
+/** @deprecated Use GeminiPricing instead. */
 export const GeminiTextPricing = GeminiPricing;
+
+// ─── Catalog ──────────────────────────────────────────────────────────────────
+// Metadata (deprecated, deprecationDate, limitedAvailability, replacedBy) is
+// read directly from the GeminiPricing entries — no need for separate groups.
+
+export const GOOGLE_MODEL_CATALOG: ModelCatalogEntry[] = [
+  ...completions([...GoogleTextModels, ...GoogleLimitedModels, ...GoogleAllDeprecatedModels], "google", GeminiPricing),
+  ...liveApi(GoogleLiveApiModels,      "google", GeminiPricing),
+  ...embeddings(GoogleEmbeddingModelsList, "google", GeminiPricing),
+  ...images(GoogleImageModels,         "google", GeminiPricing),
+  ...videos(GoogleVideoModels,         "google", GeminiPricing),
+  ...completions(GemmaTextModels,      "google", GemmaPricing),
+  ...audios(GoogleTTSModels,           "google", GeminiPricing),
+];

package/src/clients/pricing/groq.ts

@@ -0,0 +1,56 @@
+/**
+ * Groq pricing (USD per 1M tokens)
+ * Source: https://groq.com/pricing
+ */
+export const GroqTextPricing: Record<string, { input: number; output: number; cached_input?: number }> = {
+  // Llama 4
+  "meta-llama/llama-4-scout-17b-16e-instruct": { input: 0.11, output: 0.34 },
+  "meta-llama/llama-4-maverick-17b-128e-instruct": { input: 0.20, output: 0.60 },
+  // Llama Guard / Prompt Guard
+  "llama-guard-3-8b": { input: 0.20, output: 0.20 },
+  "meta-llama/llama-guard-4-12b": { input: 0.20, output: 0.20 },
+  "meta-llama/llama-prompt-guard-2-22m": { input: 0.03, output: 0.03 },
+  "meta-llama/llama-prompt-guard-2-86m": { input: 0.04, output: 0.04 },
+  // Llama 3.3
+  "llama-3.3-70b-versatile": { input: 0.59, output: 0.79 },
+  "llama-3.3-70b-specdec": { input: 0.59, output: 0.99 },
+  // Llama 3.1
+  "llama-3.1-8b-instant": { input: 0.05, output: 0.08 },
+  "llama3-70b-8192": { input: 0.59, output: 0.79 },
+  "llama3-8b-8192": { input: 0.05, output: 0.08 },
+  // DeepSeek
+  "deepseek-r1-distill-llama-70b": { input: 0.75, output: 0.99 },
+  // Gemma
+  "gemma2-9b-it": { input: 0.20, output: 0.20 },
+  // Mistral
+  "mistral-saba-24b": { input: 0.79, output: 0.79 },
+  // Compound
+  "compound-beta": { input: 0.0, output: 0.0 },
+  // Groq compound (newer aliases)
+  "groq/compound": { input: 0.0, output: 0.0 },
+  "groq/compound-mini": { input: 0.0, output: 0.0 },
+  "compound-beta-mini": { input: 0.0, output: 0.0 },
+  // Allam (free)
+  "allam-2-7b": { input: 0.0, output: 0.0 },
+  // Qwen
+  "qwen-qwq-32b": { input: 0.29, output: 0.39 },
+  "qwen-2.5-32b": { input: 0.79, output: 0.79 },
+  "qwen-2.5-coder-32b": { input: 0.79, output: 0.79 },
+  "qwen/qwen3-32b": { input: 0.29, output: 0.59 },
+  // MoonshotAI
+  "moonshotai/kimi-k2-instruct": { input: 1.0, output: 3.0 },
+  "moonshotai/kimi-k2-instruct-0905": { input: 1.0, output: 3.0 },
+  // OpenAI OSS models on Groq
+  "openai/gpt-oss-120b": { input: 0.15, output: 0.60 },
+  "openai/gpt-oss-20b": { input: 0.075, output: 0.30 },
+  "openai/gpt-oss-safeguard-20b": { input: 0.075, output: 0.30 },
+  // Canopy Labs Orpheus (TTS/speech)
+  "canopylabs/orpheus-arabic-saudi": { input: 40.0, output: 0.0 },
+  "canopylabs/orpheus-v1-english": { input: 0.0, output: 0.0 },
+  // Audio / TTS (free)
+  "whisper-large-v3": { input: 0.0, output: 0.0 },
+  "whisper-large-v3-turbo": { input: 0.0, output: 0.0 },
+  "distil-whisper-large-v3-en": { input: 0.0, output: 0.0 },
+  "playai-tts": { input: 0.0, output: 0.0 },
+  "playai-tts-arabic": { input: 0.0, output: 0.0 },
+};

package/src/clients/pricing/index.ts

@@ -1,6 +1,43 @@
-export { OpenAiTextPricing } from "./openai";
-export { GeminiTextPricing } from "./google";
-export { AnthropicTextPricing } from "./anthropic";
-export { XaiTextPricing, XaiImagePricing, XaiVideoPricing } from "./xai";
-export { ALL_MODEL_CATALOG, OPENAI_MODEL_CATALOG, ANTHROPIC_MODEL_CATALOG, GOOGLE_MODEL_CATALOG, XAI_MODEL_CATALOG, USAGE_MARKUP_PERCENT } from "./catalog";
-export type { ModelCatalogEntry, ModelType, ModelPricing as CatalogModelPricing } from "./catalog";
+// ─── Per-provider model IDs, pricing, and catalogs ───────────────────────────
+export {
+  OpenAiModels, OpenAiEmbeddingModels,
+  OpenAiTextPricing, OPENAI_MODEL_CATALOG,
+  OpenAiReasoningModels, OpenAiChatModels, OpenAiEmbeddingModelsList,
+  OpenAiResponsesOnlyModels, OpenAiImageModels, OpenAiVideoModels,
+  OpenAiTTSModels, OpenAiTranscriptionModels, OpenAiRealtimeModels,
+} from "./openai";
+
+export {
+  AnthropicModels,
+  AnthropicTextPricing, ANTHROPIC_MODEL_CATALOG,
+} from "./anthropic";
+
+export {
+  GoogleModels, GoogleEmbeddingModels,
+  GeminiPricing, GeminiTextPricing, GOOGLE_MODEL_CATALOG,
+  GoogleTextModels, GoogleImageModels, GoogleVideoModels,
+  GoogleTTSModels, GoogleEmbeddingModelsList,
+} from "./google";
+
+export {
+  XaiModels,
+  XaiTextPricing, XaiImagePricing, XaiVideoPricing, XAI_MODEL_CATALOG,
+  XaiTextModels, XaiImageModels, XaiVideoModels,
+} from "./xai";
+
+// ─── Other provider pricing ───────────────────────────────────────────────────
+export { GroqTextPricing } from "./groq";
+export { DeepSeekTextPricing } from "./deepseek";
+export { CerebrasTextPricing } from "./cerebras";
+export { MistralTextPricing } from "./mistral";
+export { NvidiaTextPricing, NvidiaImagePricing } from "./nvidia";
+export { GitHubModelsTextPricing } from "./github";
+export { OpenRouterTextPricing } from "./openrouter";
+export { LlamaTextPricing } from "./llama";
+export { CopilotTextPricing, CopilotModelMultipliers } from "./copilot";
+export { FireworksTextPricing } from "./fireworks";
+
+// ─── Combined catalog ─────────────────────────────────────────────────────────
+export { ALL_MODEL_CATALOG, USAGE_MARKUP_PERCENT } from "./models";
+export type { ModelCatalogEntry, ModelType, ModelPricing as CatalogModelPricing } from "./types";
+export { completions, embeddings, images, videos, audios, transactions, liveApi } from "./types";

package/src/clients/pricing/llama.ts

@@ -0,0 +1,18 @@
+/**
+ * Meta Llama API pricing (USD per 1M tokens)
+ * Source: https://llama.developer.meta.com/docs/
+ * Free models available directly from Meta.
+ */
+export const LlamaTextPricing: Record<string, { input: number; output: number }> = {
+  // Free models
+  "llama-3.3-70b-instruct": { input: 0.0, output: 0.0 },
+  "llama-3.3-8b-instruct": { input: 0.0, output: 0.0 },
+  "llama-4-scout-17b-16e-instruct": { input: 0.0, output: 0.0 },
+  "llama-4-maverick-17b-128e-instruct": { input: 0.0, output: 0.0 },
+  // Cerebras-hosted Llama 4 (free)
+  "cerebras-llama-4-maverick-17b-128e-instruct": { input: 0.0, output: 0.0 },
+  "cerebras-llama-4-scout-17b-16e-instruct": { input: 0.0, output: 0.0 },
+  // Groq-hosted Llama 4 (free)
+  "groq-llama-4-maverick-17b-128e-instruct": { input: 0.0, output: 0.0 },
+  "groq-llama-4-scout-17b-16e-instruct": { input: 0.0, output: 0.0 },
+};

package/src/clients/pricing/mistral.ts

@@ -0,0 +1,34 @@
+/**
+ * Mistral AI pricing (USD per 1M tokens)
+ * Source: https://mistral.ai/technology/#pricing
+ */
+export const MistralTextPricing: Record<string, { input: number; output: number; cached_input?: number }> = {
+  // Mistral Large (latest = 2512 as of 2026-04, price reduced from $2/$6)
+  // Source: openrouter.ai/api/v1/models (mistral-large-2512 = $0.5/$1.5)
+  "mistral-large-latest": { input: 0.5, output: 1.5 },
+  "mistral-large-2512": { input: 0.5, output: 1.5 },
+  "mistral-large-2411": { input: 2.0, output: 6.0 },
+  // Mistral Small (latest = 2603 as of 2026-04, price updated)
+  // Source: openrouter.ai/api/v1/models (mistral-small-2603 = $0.15/$0.60)
+  "mistral-small-latest": { input: 0.15, output: 0.6 },
+  "mistral-small-2603": { input: 0.15, output: 0.6 },
+  "mistral-small-2501": { input: 0.1, output: 0.3 },
+  // Mistral Medium
+  "mistral-medium-latest": { input: 0.4, output: 2.0 },
+  // Codestral (code model)
+  "codestral-latest": { input: 0.3, output: 0.9 },
+  "codestral-2501": { input: 0.3, output: 0.9 },
+  // Devstral (free coding assistant)
+  "devstral-small-latest": { input: 0.0, output: 0.0 },
+  "labs-devstral-small-2512": { input: 0.0, output: 0.0 },
+  // Pixtral (multimodal)
+  "pixtral-large-latest": { input: 2.0, output: 6.0 },
+  "pixtral-12b-2409": { input: 0.15, output: 0.15 },
+  // Ministral
+  "ministral-3b-latest": { input: 0.04, output: 0.04 },
+  "ministral-8b-latest": { input: 0.1, output: 0.1 },
+  // Mistral NeMo
+  "open-mistral-nemo": { input: 0.15, output: 0.15 },
+  // Embeddings
+  "mistral-embed": { input: 0.1, output: 0.0 },
+};

package/src/clients/pricing/models.ts

@@ -0,0 +1,23 @@
+/**
+ * Assembles ALL_MODEL_CATALOG from per-provider catalogs.
+ * Each provider file is the single source of truth for its models.
+ */
+import { ModelPricing, ModelType, ModelCatalogEntry } from "./types";
+import { OPENAI_MODEL_CATALOG } from "./openai";
+import { ANTHROPIC_MODEL_CATALOG } from "./anthropic";
+import { GOOGLE_MODEL_CATALOG } from "./google";
+import { XAI_MODEL_CATALOG } from "./xai";
+
+export { ModelPricing, ModelType, ModelCatalogEntry };
+
+/** 2.5% platform markup applied on top of all provider base rates */
+export const USAGE_MARKUP_PERCENT = 2.5 / 100;
+
+export { OPENAI_MODEL_CATALOG, ANTHROPIC_MODEL_CATALOG, GOOGLE_MODEL_CATALOG, XAI_MODEL_CATALOG };
+
+export const ALL_MODEL_CATALOG: ModelCatalogEntry[] = [
+  ...OPENAI_MODEL_CATALOG,
+  ...ANTHROPIC_MODEL_CATALOG,
+  ...GOOGLE_MODEL_CATALOG,
+  ...XAI_MODEL_CATALOG,
+];