@shop-os/modelkit 0.1.7 → 0.1.9

package/dist/index.js

@@ -35,6 +35,23 @@ var Catalog = class _Catalog {
     for (const model of models) this.add(model);
   }
   add(model) {
+    if (this.byId.has(model.id)) {
+      throw new Error(`Duplicate model id "${model.id}"`);
+    }
+    if (this.byAlias.has(model.id)) {
+      throw new Error(
+        `Model id "${model.id}" collides with an alias of "${this.byAlias.get(model.id)}"`
+      );
+    }
+    for (const alias of model.aliases) {
+      if (alias === model.id) {
+        throw new Error(`Alias "${alias}" collides with its own model id "${model.id}"`);
+      }
+      const existing = this.byAlias.get(alias);
+      if (existing !== void 0) {
+        throw new Error(`Duplicate alias "${alias}" (already maps to "${existing}")`);
+      }
+    }
     this.byId.set(model.id, model);
     this.byAlias.set(model.id, model.id);
     for (const alias of model.aliases) this.byAlias.set(alias, model.id);
@@ -97,7 +114,7 @@ var GEMINI_RATIOS = [
   "5:4",
   "adaptive"
 ];
-function geminiImage(id, label, aliases, sizes, display) {
+function geminiImage(id, label, aliases, sizes, display, exposure) {
   const model = {
     id,
     aliases,
@@ -106,13 +123,14 @@ function geminiImage(id, label, aliases, sizes, display) {
     adapter: "google-image",
     providerModelId: id,
     lifecycle: "active",
+    exposure,
     constraints: {
       aspectRatios: [...GEMINI_RATIOS],
       sizes,
       supports: ["safetySettings", "temperature"]
     },
     defaults: { size: "1K", aspectRatio: "1:1" },
-    usage: { ...IMAGE_USAGE },
+    usage: { ...IMAGE_USAGE, slug: "gemini" },
     pricingKind: "image_size",
     display,
     capabilities: { ...IMAGE_GENERATE_EDIT }
@@ -120,14 +138,21 @@ function geminiImage(id, label, aliases, sizes, display) {
   return model;
 }
 var IMAGE_MODELS = [
-  geminiImage("gemini-3-pro-image-preview", "Nano Banana Pro", ["nb-pro"], ["1K", "2K", "4K"], {
-    provider: "google",
-    description: "Create stunning, production-quality visuals with exceptional realism, detail, and typography. Best for campaigns, e-commerce, product photography, and hero creatives.",
-    logo: "https://d31vvfpk2ppdyq.cloudfront.net/assets/image/192485d2-a76a-4657-9c06-930b2c70badd_compressed.png"
-  }),
+  geminiImage(
+    "gemini-3-pro-image-preview",
+    "Nano Banana Pro",
+    ["nb-pro"],
+    ["1K", "2K", "4K"],
+    {
+      provider: "google",
+      description: "Create stunning, production-quality visuals with exceptional realism, detail, and typography. Best for campaigns, e-commerce, product photography, and hero creatives.",
+      logo: "https://d31vvfpk2ppdyq.cloudfront.net/assets/image/192485d2-a76a-4657-9c06-930b2c70badd_compressed.png"
+    },
+    "catalog"
+  ),
   geminiImage(
     "gemini-3.1-flash-image-preview",
-    "Nano Banana Flash",
+    "Nano Banana 2",
     ["nb-2"],
     // Flash 3.1 is the only Gemini image model with the 512 tier.
     ["512", "1K", "2K", "4K"],
@@ -135,35 +160,61 @@ var IMAGE_MODELS = [
       provider: "google",
       description: "Generate high-quality images at speed. Perfect for brainstorming, moodboards, concept testing, creative exploration, and large batch workflows.",
       logo: "https://d31vvfpk2ppdyq.cloudfront.net/assets/image/192485d2-a76a-4657-9c06-930b2c70badd_compressed.png"
-    }
+    },
+    "catalog"
+  ),
+  geminiImage(
+    "gemini-3.1-flash-lite-image",
+    "Nano Banana Lite",
+    ["nb-lite"],
+    ["1K"],
+    {
+      provider: "google",
+      description: "The fastest and most affordable image model. Ideal for rapid drafts, bulk exploration, and iterative concept testing at 1K resolution.",
+      logo: "https://d31vvfpk2ppdyq.cloudfront.net/assets/image/192485d2-a76a-4657-9c06-930b2c70badd_compressed.png"
+    },
+    "catalog"
   ),
-  geminiImage("gemini-3.1-flash-lite-image", "Nano Banana Lite", ["nb-lite"], ["1K"], {
-    provider: "google",
-    description: "The fastest and most affordable image model. Ideal for rapid drafts, bulk exploration, and iterative concept testing at 1K resolution.",
-    logo: "https://d31vvfpk2ppdyq.cloudfront.net/assets/image/192485d2-a76a-4657-9c06-930b2c70badd_compressed.png"
-  }),
   // trigger-workflows validates 1K only; pricing seeds 2K/4K rows so allow all three.
-  geminiImage("gemini-2.5-flash-image", "Gemini 2.5 Flash Image", [], ["1K", "2K", "4K"], {
-    provider: "google",
-    description: "Nano Banana (legacy Gemini 2.5 Flash Image) native image generation and editing model for high-volume, low-latency creative workflows."
-  }),
+  geminiImage(
+    "gemini-2.5-flash-image",
+    "Gemini 2.5 Flash Image",
+    [],
+    ["1K", "2K", "4K"],
+    {
+      provider: "google",
+      description: "Nano Banana (legacy Gemini 2.5 Flash Image) native image generation and editing model for high-volume, low-latency creative workflows."
+    },
+    "internal"
+  ),
   {
     // Cowork "auto" lets OpenAI pick the size; the canonical equivalent is adaptive.
     id: "gpt-image-2",
     aliases: [],
-    label: "GPT Image 2",
+    label: "GPT-image-2",
     type: "image",
     adapter: "openai-image",
     providerModelId: "gpt-image-2",
     lifecycle: "active",
+    exposure: "catalog",
     constraints: {
+      exactSizes: {
+        "1:1": "1024x1024",
+        "3:2": "1536x1024",
+        "2:3": "1024x1536",
+        "16:9": "2048x1152",
+        "9:16": "1152x2048",
+        "4:5": "1024x1280",
+        "5:4": "1280x1024"
+      },
       aspectRatios: ["1:1", "3:2", "2:3", "16:9", "9:16", "4:5", "5:4", "adaptive"],
       sizes: ["1K", "2K", "4K"],
+      qualityOptions: ["auto", "low", "medium", "high"],
       references: { images: 16 },
       supports: ["quality", "count"]
     },
     defaults: { size: "1K", aspectRatio: "adaptive", quality: "auto" },
-    usage: { ...IMAGE_USAGE },
+    usage: { ...IMAGE_USAGE, slug: "gpt" },
     pricingKind: "image_size",
     display: {
       provider: "openai",
@@ -181,9 +232,20 @@ var IMAGE_MODELS = [
     adapter: "openai-image",
     providerModelId: "gpt-image-2.5-sunburst",
     lifecycle: "active",
+    exposure: "internal",
     constraints: {
+      exactSizes: {
+        "1:1": "1024x1024",
+        "3:2": "1536x1024",
+        "2:3": "1024x1536",
+        "16:9": "2048x1152",
+        "9:16": "1152x2048",
+        "4:5": "1024x1280",
+        "5:4": "1280x1024"
+      },
       aspectRatios: ["1:1", "3:2", "2:3", "16:9", "9:16", "4:5", "5:4", "adaptive"],
       sizes: ["1K", "2K", "4K"],
+      qualityOptions: ["auto", "low", "medium", "high"],
       references: { images: 16 },
       supports: ["quality", "count"]
     },
@@ -205,9 +267,20 @@ var IMAGE_MODELS = [
     adapter: "openai-image",
     providerModelId: "gpt-image-2.5-flare",
     lifecycle: "active",
+    exposure: "internal",
     constraints: {
+      exactSizes: {
+        "1:1": "1024x1024",
+        "3:2": "1536x1024",
+        "2:3": "1024x1536",
+        "16:9": "2048x1152",
+        "9:16": "1152x2048",
+        "4:5": "1024x1280",
+        "5:4": "1280x1024"
+      },
       aspectRatios: ["1:1", "3:2", "2:3", "16:9", "9:16", "4:5", "5:4", "adaptive"],
       sizes: ["1K", "2K", "4K"],
+      qualityOptions: ["auto", "low", "medium", "high"],
       references: { images: 16 },
       supports: ["quality", "count"]
     },
@@ -223,12 +296,13 @@ var IMAGE_MODELS = [
   {
     // Cowork allows 7 ratios; trigger-workflows allows 9 (adds 3:2, 9:21, drops 4:5/5:4).
     id: "muse-image-1.0",
-    aliases: [],
+    aliases: ["meta-muse"],
     label: "Meta Muse",
     type: "image",
     adapter: "fal",
     providerModelId: "muse-image-1.0",
     lifecycle: "active",
+    exposure: "catalog",
     constraints: {
       aspectRatios: [
         "21:9",
@@ -249,7 +323,7 @@ var IMAGE_MODELS = [
       supports: ["count", "outputFormat"]
     },
     defaults: { size: "1K", aspectRatio: "1:1", outputFormat: "png", count: 1 },
-    usage: { ...IMAGE_USAGE },
+    usage: { ...IMAGE_USAGE, slug: "muse" },
     pricingKind: "image_size",
     display: {
       provider: "meta",
@@ -266,6 +340,7 @@ var IMAGE_MODELS = [
     adapter: "ark",
     providerModelId: "dola-seedream-5-0-pro-260628",
     lifecycle: "active",
+    exposure: "catalog",
     constraints: {
       aspectRatios: ["1:1", "4:3", "3:4", "16:9", "9:16", "3:2", "2:3", "21:9", "4:5", "5:4"],
       sizes: ["1K", "2K"],
@@ -291,6 +366,7 @@ var IMAGE_MODELS = [
     adapter: "ark",
     providerModelId: "seedream-4-5-251128",
     lifecycle: "active",
+    exposure: "internal",
     constraints: {
       // 97:30 has no Ark size spelling; the A+ flow pads inputs to 97:30 and
       // omits the ratio instead. Advertise only sizes the adapter can send.
@@ -328,6 +404,7 @@ var IMAGE_MODELS = [
     adapter: "prodia",
     providerModelId: "flux-2-pro",
     lifecycle: "active",
+    exposure: "catalog",
     constraints: {
       aspectRatios: ["1:1", "2:3", "3:2", "16:9", "9:16", "4:3", "3:4", "4:5", "5:4", "21:9"],
       sizes: ["1K"],
@@ -354,6 +431,7 @@ var IMAGE_MODELS = [
     adapter: "prodia",
     providerModelId: "inference.gemini-3-pro.img2img.v1",
     lifecycle: "active",
+    exposure: "internal",
     constraints: {
       aspectRatios: ["1:1", "2:3", "3:2", "4:3", "3:4", "4:5", "5:4", "9:16", "16:9", "21:9"],
       sizes: ["1K", "2K", "4K"],
@@ -379,6 +457,7 @@ var IMAGE_MODELS = [
     adapter: "prodia",
     providerModelId: "inference.nano-banana.img2img.v2",
     lifecycle: "active",
+    exposure: "internal",
     constraints: {
       aspectRatios: ["1:1", "2:3", "3:4", "4:3", "9:16", "16:9", "21:9"],
       sizes: ["1K"],
@@ -405,6 +484,7 @@ var IMAGE_MODELS = [
     adapter: "replicate",
     providerModelId: "wan-video/wan-2.7-image-pro",
     lifecycle: "active",
+    exposure: "internal",
     constraints: {
       aspectRatios: ["1:1", "16:9", "9:16", "4:3", "3:4"],
       sizes: ["1K", "2K", "4K"],
@@ -431,6 +511,7 @@ var IMAGE_MODELS = [
     adapter: "replicate",
     providerModelId: "qwen/qwen-image-2-pro",
     lifecycle: "active",
+    exposure: "internal",
     constraints: {
       aspectRatios: ["1:1", "16:9", "9:16", "4:3", "3:4", "3:2", "2:3"],
       sizes: [],
@@ -458,6 +539,7 @@ var IMAGE_MODELS = [
     adapter: "replicate",
     providerModelId: "qwen-edit-apps/qwen-image-edit-plus-lora-skin",
     lifecycle: "active",
+    exposure: "internal",
     constraints: {
       aspectRatios: ["1:1", "16:9", "9:16", "4:3", "3:4"],
       sizes: [],
@@ -486,6 +568,7 @@ var OTHER_MODELS = [
     adapter: "google-genai",
     providerModelId: "lyria-3-clip-preview",
     lifecycle: "active",
+    exposure: "catalog",
     constraints: {
       durations: [30],
       references: { images: 10 },
@@ -497,7 +580,8 @@ var OTHER_MODELS = [
     display: {
       provider: "google",
       description: "Google Lyria 3 Clip short music clips, loops, and previews; takes text and up to 10 image prompts and returns a fixed 30-second MP3 plus lyrics."
-    }
+    },
+    capabilities: {}
   },
   {
     id: "lyria-3-pro-preview",
@@ -507,6 +591,7 @@ var OTHER_MODELS = [
     adapter: "google-genai",
     providerModelId: "lyria-3-pro-preview",
     lifecycle: "active",
+    exposure: "catalog",
     constraints: {
       references: { images: 10 },
       supports: []
@@ -517,7 +602,8 @@ var OTHER_MODELS = [
     display: {
       provider: "google",
       description: "Google Lyria 3 Pro flagship music model for full-length songs with verses, choruses, and bridges; takes text and up to 10 image prompts and returns MP3 plus lyrics."
-    }
+    },
+    capabilities: {}
   },
   {
     id: "google/lyria-2",
@@ -527,6 +613,7 @@ var OTHER_MODELS = [
     adapter: "replicate",
     providerModelId: "google/lyria-2",
     lifecycle: "active",
+    exposure: "internal",
     constraints: {
       durations: [30],
       supports: ["seed", "negativePrompt"]
@@ -537,7 +624,8 @@ var OTHER_MODELS = [
     display: {
       provider: "google",
       description: "Lyria 2 music generation on Replicate; produces 30 seconds of 48 kHz stereo audio from a text prompt."
-    }
+    },
+    capabilities: {}
   },
   {
     id: "mmaudio",
@@ -547,6 +635,7 @@ var OTHER_MODELS = [
     adapter: "replicate",
     providerModelId: "zsxkib/mmaudio:62871fb59889b2d7c13777f08deb3b36bdff88f7e1d53a50ad7694548a41b484",
     lifecycle: "active",
+    exposure: "catalog",
     constraints: {
       durations: { min: 1, max: 30 },
       requires: ["sourceVideo"],
@@ -558,7 +647,8 @@ var OTHER_MODELS = [
     display: {
       provider: "mmaudio",
       description: "MMAudio V2 video-to-audio model on Replicate; takes a source video and prompt and synthesizes synchronized environmental, action, and speech-like audio."
-    }
+    },
+    capabilities: {}
   },
   {
     id: "elevenlabs-music",
@@ -568,6 +658,7 @@ var OTHER_MODELS = [
     adapter: "replicate",
     providerModelId: "elevenlabs/music",
     lifecycle: "active",
+    exposure: "internal",
     constraints: {
       durations: { min: 5, max: 300 },
       supports: ["instrumental"]
@@ -579,7 +670,8 @@ var OTHER_MODELS = [
     display: {
       provider: "elevenlabs",
       description: "ElevenLabs Music text-to-music on Replicate; generates full songs, instrumental or with vocals, up to 5 minutes long at up to 44.1 kHz."
-    }
+    },
+    capabilities: {}
   },
   {
     id: "topaz",
@@ -589,6 +681,7 @@ var OTHER_MODELS = [
     adapter: "replicate",
     providerModelId: "topazlabs/image-upscale",
     lifecycle: "active",
+    exposure: "catalog",
     constraints: {
       references: { images: 1 },
       requires: ["references"],
@@ -611,6 +704,7 @@ var OTHER_MODELS = [
     adapter: "prodia",
     providerModelId: "inference.remove-background.v1",
     lifecycle: "active",
+    exposure: "catalog",
     constraints: {
       references: { images: 1 },
       requires: ["references"],
@@ -638,6 +732,7 @@ function text(id, label, providerModelId, extra) {
     adapter: "ai-sdk-text",
     providerModelId,
     lifecycle: "active",
+    exposure: "internal",
     constraints: { supports: [] },
     defaults: {},
     usage: { ...TEXT_USAGE },
@@ -645,39 +740,57 @@ function text(id, label, providerModelId, extra) {
     ...extra
   };
 }
+var ANTHROPIC_THINKING = {
+  anthropic: { thinking: { type: "enabled", budgetTokens: 4e3 }, sendReasoning: true }
+};
+var ANTHROPIC_ADAPTIVE = {
+  anthropic: { thinking: { type: "adaptive", display: "summarized" }, sendReasoning: true }
+};
+var OPENAI_REASONING = { openai: { reasoningEffort: "medium", reasoningSummary: "auto" } };
+var DEEPSEEK_THINKING = { deepseek: { thinking: { type: "enabled" } } };
 var TEXT_MODELS = [
   text("gpt-4o", "GPT-4o", "gpt-4o", {
+    exposure: "internal",
     display: {
       provider: "openai",
+      costTier: 2,
       description: "OpenAI GPT-4o flagship multimodal model; text and image input, 128K-token context."
     },
     capabilities: { vision: true, tools: true, structuredOutput: true, contextWindow: 128e3 }
   }),
   text("gpt-4o-mini", "GPT-4o mini", "gpt-4o-mini", {
+    exposure: "internal",
     display: {
       provider: "openai",
+      costTier: 1,
       description: "OpenAI GPT-4o mini fast, low-cost multimodal model; text and image input, 128K-token context."
     },
     capabilities: { vision: true, tools: true, structuredOutput: true, contextWindow: 128e3 }
   }),
   text("gpt-5.1", "GPT-5.1", "gpt-5.1", {
+    exposure: "catalog",
     display: {
       provider: "openai",
+      costTier: 3,
       description: "OpenAI GPT-5.1 reasoning model for coding and agentic tasks; text and image input, 400K-token context."
     },
     capabilities: { vision: true, tools: true, structuredOutput: true, contextWindow: 4e5 }
   }),
   text("gpt-5.2", "GPT-5.2", "gpt-5.2", {
+    exposure: "catalog",
     display: {
       provider: "openai",
+      costTier: 3,
       description: "OpenAI GPT-5.2 reasoning model for complex professional work; text and image input, 400K-token context."
     },
     capabilities: { vision: true, tools: true, structuredOutput: true, contextWindow: 4e5 }
   }),
   // 1.05M-token context: https://platform.openai.com/docs/models/gpt-5.4
   text("gpt-5.4", "GPT-5.4", "gpt-5.4", {
+    exposure: "internal",
     display: {
       provider: "openai",
+      costTier: 3,
       description: "OpenAI GPT-5.4 flagship reasoning model; text and image input, 1.05M-token context."
     },
     capabilities: {
@@ -689,16 +802,20 @@ var TEXT_MODELS = [
   }),
   // OpenAI never shipped gpt-5.1-mini or gpt-5.2-mini; callers that still send them get 5.4 mini.
   text("gpt-5.4-mini", "GPT-5.4 mini", "gpt-5.4-mini", {
+    exposure: "catalog",
     aliases: ["gpt-5.1-mini", "gpt-5.2-mini"],
     display: {
       provider: "openai",
+      costTier: 1,
       description: "OpenAI GPT-5.4 mini efficient reasoning model for high-volume workloads; text and image input, 400K-token context."
     },
     capabilities: { vision: true, tools: true, structuredOutput: true, contextWindow: 4e5 }
   }),
   text("gpt-5.5", "GPT-5.5", "gpt-5.5", {
+    exposure: "catalog",
     display: {
       provider: "openai",
+      costTier: 3,
       description: "OpenAI GPT-5.5 flagship reasoning model for the most complex professional and coding work; text and image input, 1.05M-token context."
     },
     capabilities: {
@@ -709,11 +826,15 @@ var TEXT_MODELS = [
     }
   }),
   text("gpt-5.6-luna", "GPT-5.6 Luna", "gpt-5.6-luna", {
+    providerOptions: OPENAI_REASONING,
+    exposure: "catalog",
     display: {
       provider: "openai",
+      costTier: 1,
       description: "OpenAI GPT-5.6 Luna cost-optimized tier of the GPT-5.6 reasoning family; text and image input, 1.05M-token context."
     },
     capabilities: {
+      reasoning: true,
       vision: true,
       tools: true,
       structuredOutput: true,
@@ -721,11 +842,15 @@ var TEXT_MODELS = [
     }
   }),
   text("gpt-5.6-terra", "GPT-5.6 Terra", "gpt-5.6-terra", {
+    providerOptions: OPENAI_REASONING,
+    exposure: "catalog",
     display: {
       provider: "openai",
+      costTier: 2,
       description: "OpenAI GPT-5.6 Terra balanced tier of the GPT-5.6 reasoning family; text and image input, 1.05M-token context."
     },
     capabilities: {
+      reasoning: true,
       vision: true,
       tools: true,
       structuredOutput: true,
@@ -733,11 +858,15 @@ var TEXT_MODELS = [
     }
   }),
   text("gpt-5.6-sol", "GPT-5.6 Sol", "gpt-5.6-sol", {
+    providerOptions: OPENAI_REASONING,
+    exposure: "catalog",
     display: {
       provider: "openai",
+      costTier: 3,
       description: "OpenAI GPT-5.6 Sol flagship tier of the GPT-5.6 reasoning family; text and image input, 1.05M-token context."
     },
     capabilities: {
+      reasoning: true,
       vision: true,
       tools: true,
       structuredOutput: true,
@@ -746,15 +875,19 @@ var TEXT_MODELS = [
   }),
   // Library key "claude-sonnet-4.5" maps to the SDK id "claude-sonnet-4-5".
   text("claude-sonnet-4-5", "Claude Sonnet 4.5", "claude-sonnet-4-5", {
+    exposure: "internal",
     aliases: ["claude-sonnet-4.5", "claude-sonnet-4-5-20250929"],
     display: {
       provider: "anthropic",
+      costTier: 2,
       description: "Anthropic Claude Sonnet 4.5 (legacy) balanced model; text and image input, 200K-token context."
     },
     capabilities: { vision: true, tools: true, structuredOutput: true, contextWindow: 2e5 }
   }),
   // Anthropic retired the dated 4.0 ids; callers that still send them get 4.6.
   text("claude-sonnet-4-6", "Claude Sonnet 4.6", "claude-sonnet-4-6", {
+    providerOptions: ANTHROPIC_THINKING,
+    exposure: "catalog",
     aliases: [
       "claude-sonnet-4.6",
       "claude-sonnet-4",
@@ -763,9 +896,11 @@ var TEXT_MODELS = [
     ],
     display: {
       provider: "anthropic",
+      costTier: 2,
       description: "Anthropic Claude Sonnet 4.6 (legacy) balanced model; text and image input, 1M-token context."
     },
     capabilities: {
+      reasoning: true,
       vision: true,
       tools: true,
       structuredOutput: true,
@@ -773,11 +908,15 @@ var TEXT_MODELS = [
     }
   }),
   text("claude-sonnet-5", "Claude Sonnet 5", "claude-sonnet-5", {
+    providerOptions: ANTHROPIC_ADAPTIVE,
+    exposure: "catalog",
     display: {
       provider: "anthropic",
+      costTier: 3,
       description: "Anthropic Claude Sonnet 5 current balanced model; text and image input, 1M-token context."
     },
     capabilities: {
+      reasoning: true,
       vision: true,
       tools: true,
       structuredOutput: true,
@@ -785,12 +924,16 @@ var TEXT_MODELS = [
     }
   }),
   text("claude-opus-4-8", "Claude Opus 4.8", "claude-opus-4-8", {
+    providerOptions: ANTHROPIC_ADAPTIVE,
+    exposure: "catalog",
     aliases: ["claude-opus-4.8", "claude-opus-4", "claude-opus-4.0", "claude-opus-4-20250514"],
     display: {
       provider: "anthropic",
+      costTier: 3,
       description: "Anthropic Claude Opus 4.8 (legacy) frontier model; text and image input, 1M-token context."
     },
     capabilities: {
+      reasoning: true,
       vision: true,
       tools: true,
       structuredOutput: true,
@@ -798,17 +941,28 @@ var TEXT_MODELS = [
     }
   }),
   text("claude-haiku-4-5", "Claude Haiku 4.5", "claude-haiku-4-5", {
+    providerOptions: ANTHROPIC_THINKING,
+    exposure: "catalog",
     aliases: ["claude-haiku-4.5", "claude-haiku-4-5-20251001", "claude-haiku-4"],
     display: {
       provider: "anthropic",
+      costTier: 1,
       description: "Anthropic Claude Haiku 4.5 fast model with near-frontier intelligence; text and image input, 200K-token context."
     },
-    capabilities: { vision: true, tools: true, structuredOutput: true, contextWindow: 2e5 }
+    capabilities: {
+      reasoning: true,
+      vision: true,
+      tools: true,
+      structuredOutput: true,
+      contextWindow: 2e5
+    }
   }),
   text("claude-opus-5", "Claude Opus 5", "claude-opus-5", {
+    exposure: "internal",
     aliases: ["claude-opus-5.0"],
     display: {
       provider: "anthropic",
+      costTier: 3,
       description: "Anthropic Claude Opus 5 current frontier model for agentic coding and enterprise work; text and image input, 1M-token context."
     },
     capabilities: {
@@ -819,8 +973,10 @@ var TEXT_MODELS = [
     }
   }),
   text("claude-fable-5", "Claude Fable 5", "claude-fable-5", {
+    exposure: "internal",
     display: {
       provider: "anthropic",
+      costTier: 3,
       description: "Anthropic Claude Fable 5 (legacy) model for demanding reasoning and long-horizon agentic work; text and image input, 1M-token context."
     },
     capabilities: {
@@ -831,8 +987,10 @@ var TEXT_MODELS = [
     }
   }),
   text("gemini-2.5-flash", "Gemini 2.5 Flash", "gemini-2.5-flash", {
+    exposure: "internal",
     display: {
       provider: "google",
+      costTier: 1,
       description: "Google Gemini 2.5 Flash low-latency multimodal model for high-volume reasoning over text, images, video, and audio; 1M-token context."
     },
     capabilities: {
@@ -843,8 +1001,10 @@ var TEXT_MODELS = [
     }
   }),
   text("gemini-2.5-pro", "Gemini 2.5 Pro", "gemini-2.5-pro", {
+    exposure: "internal",
     display: {
       provider: "google",
+      costTier: 2,
       description: "Google Gemini 2.5 Pro thinking model for complex reasoning over code, math, and documents; text, image, video, audio, and PDF input, 1M-token context."
     },
     capabilities: {
@@ -855,8 +1015,10 @@ var TEXT_MODELS = [
     }
   }),
   text("gemini-3.1-pro-preview", "Gemini 3.1 Pro (preview)", "gemini-3.1-pro-preview", {
+    exposure: "catalog",
     display: {
       provider: "google",
+      costTier: 2,
       description: "Google Gemini 3.1 Pro preview reasoning model for software engineering, tool use, and reliable agentic workflows; 1M-token context."
     },
     capabilities: {
@@ -867,8 +1029,10 @@ var TEXT_MODELS = [
     }
   }),
   text("gemini-3.5-flash", "Gemini 3.5 Flash", "gemini-3.5-flash", {
+    exposure: "catalog",
     display: {
       provider: "google",
+      costTier: 1,
       description: "Google Gemini 3.5 Flash high-throughput model for frontier-level agentic tasks; text, image, video, audio, and PDF input, 1M-token context."
     },
     capabilities: {
@@ -879,8 +1043,10 @@ var TEXT_MODELS = [
     }
   }),
   text("gemini-3.6-flash", "Gemini 3.6 Flash", "gemini-3.6-flash", {
+    exposure: "catalog",
     display: {
       provider: "google",
+      costTier: 1,
       description: "Google Gemini 3.6 Flash general-purpose multimodal agentic model balancing speed and capability; 1M-token context."
     },
     capabilities: {
@@ -891,8 +1057,11 @@ var TEXT_MODELS = [
     }
   }),
   text("gemini-3.7-flash", "Gemini 3.7 Flash", "gemini-3.7-flash", {
+    exposure: "catalog",
     display: {
       provider: "google",
+      costTier: 1,
+      recommended: true,
       description: "Google Gemini 3.7 Flash natively multimodal reasoning model for coding, agentic workflows, and multi-step execution; 1M-token context."
     },
     capabilities: {
@@ -903,8 +1072,10 @@ var TEXT_MODELS = [
     }
   }),
   text("gemini-3.8-flash", "Gemini 3.8 Flash", "gemini-3.8-flash", {
+    exposure: "internal",
     display: {
       provider: "google",
+      costTier: 1,
       description: "Google Gemini 3.8 Flash frontier model for long-horizon software engineering, autonomous agents, and enterprise workflows; 1M-token context."
     },
     capabilities: {
@@ -916,12 +1087,15 @@ var TEXT_MODELS = [
   }),
   // Renamed to gemini-3.6-flash in vibe-shared model-migration.ts.
   text("gemini-3-flash-preview", "Gemini 3 Flash (preview)", "gemini-3-flash-preview", {
+    exposure: "internal",
     lifecycle: "deprecated",
     replacedBy: "gemini-3.6-flash"
   }),
   text("gemini-3.1-flash-lite", "Gemini 3.1 Flash Lite", "gemini-3.1-flash-lite", {
+    exposure: "internal",
     display: {
       provider: "google",
+      costTier: 1,
       description: "Google Gemini 3.1 Flash-Lite low-latency multimodal model for high-volume agentic, translation, and data-extraction tasks; 1M-token context."
     },
     capabilities: {
@@ -932,8 +1106,10 @@ var TEXT_MODELS = [
     }
   }),
   text("sonar", "Sonar", "sonar", {
+    exposure: "internal",
     display: {
       provider: "perplexity",
+      costTier: 1,
       description: "Perplexity Sonar lightweight web-grounded search model returning cited answers; 128K-token context."
     },
     capabilities: {
@@ -944,8 +1120,10 @@ var TEXT_MODELS = [
     }
   }),
   text("sonar-pro", "Sonar Pro", "sonar-pro", {
+    exposure: "internal",
     display: {
       provider: "perplexity",
+      costTier: 2,
       description: "Perplexity Sonar Pro advanced web-grounded search model for complex multi-step queries; 200K-token context."
     },
     capabilities: {
@@ -956,8 +1134,10 @@ var TEXT_MODELS = [
     }
   }),
   text("sonar-reasoning-pro", "Sonar Reasoning Pro", "sonar-reasoning-pro", {
+    exposure: "internal",
     display: {
       provider: "perplexity",
+      costTier: 2,
       description: "Perplexity Sonar Reasoning Pro chain-of-thought web-grounded search model for complex problem-solving; 128K-token context."
     },
     capabilities: {
@@ -967,12 +1147,21 @@ var TEXT_MODELS = [
       contextWindow: 128e3
     }
   }),
-  text("deepseek-v4-flash", "DeepSeek V4 Flash", "deepseek-v4-flash", {
+  // DeepSeek retired V4 Flash upstream and routes the old id to V4.1 Flash at
+  // Flash rates (https://api-docs.deepseek.com/news/news260910/); the old id
+  // stays an alias so stored selections keep resolving.
+  text("deepseek-flash", "DeepSeek V4.1 Flash", "deepseek-flash", {
+    providerOptions: DEEPSEEK_THINKING,
+    aliases: ["deepseek-v4-flash"],
+    exposure: "catalog",
     display: {
       provider: "deepseek",
-      description: "DeepSeek V4 Flash thinking/non-thinking model with vision, tool calls, and JSON output; 1M-token context and up to 384K output tokens."
+      costTier: 1,
+      recommended: true,
+      description: "DeepSeek V4.1 Flash thinking/non-thinking model with vision, tool calls, and JSON output; 1M-token context and up to 384K output tokens."
     },
     capabilities: {
+      reasoning: true,
       vision: true,
       tools: true,
       structuredOutput: true,
@@ -980,11 +1169,15 @@ var TEXT_MODELS = [
     }
   }),
   text("deepseek-v4-pro", "DeepSeek V4 Pro", "deepseek-v4-pro", {
+    providerOptions: DEEPSEEK_THINKING,
+    exposure: "catalog",
     display: {
       provider: "deepseek",
+      costTier: 2,
       description: "DeepSeek V4 Pro thinking/non-thinking text model with tool calls and JSON output (no vision); 1M-token context and up to 384K output tokens."
     },
     capabilities: {
+      reasoning: true,
       vision: false,
       tools: true,
       structuredOutput: true,
@@ -993,8 +1186,10 @@ var TEXT_MODELS = [
   }),
   // 1M-token context: https://docs.x.ai/docs/models/grok-4.3
   text("grok-4.3", "Grok 4.3", "grok-4.3", {
+    exposure: "catalog",
     display: {
       provider: "xai",
+      costTier: 2,
       description: "xAI Grok 4.3 reasoning model with tool calling and structured outputs; text and image input, 1M-token context."
     },
     capabilities: {
@@ -1006,18 +1201,25 @@ var TEXT_MODELS = [
   }),
   // Moonshot retired kimi-k2.5 on 2026-08-31 and no longer publishes its specs.
   text("kimi-k2.5", "Kimi K2.5", "kimi-k2.5", {
+    defaults: { temperature: 1 },
+    exposure: "catalog",
     display: {
       provider: "moonshot",
+      costTier: 2,
       description: "Moonshot Kimi K2.5; the provider retired the model on 2026-08-31 and calls now return a model-not-found error."
     },
     capabilities: {}
   }),
   text("kimi-k3", "Kimi K3", "kimi-k3", {
+    defaults: { temperature: 1 },
+    exposure: "catalog",
     display: {
       provider: "moonshot",
+      costTier: 2,
       description: "Moonshot Kimi K3 flagship model with native visual understanding and always-on thinking; text, image, and video input, 1M-token context."
     },
     capabilities: {
+      reasoning: true,
       vision: true,
       tools: true,
       structuredOutput: true,
@@ -1027,9 +1229,11 @@ var TEXT_MODELS = [
   // Cowork calls the provider "MiniMax-M3"; the pricing key is lowercase.
   // 1M-token context: https://platform.minimax.io/docs/api-reference/text-openai-api
   text("minimax-m3", "MiniMax M3", "MiniMax-M3", {
+    exposure: "catalog",
     aliases: ["MiniMax-M3"],
     display: {
       provider: "minimax",
+      costTier: 2,
       description: "MiniMax M3 frontier multimodal model for agentic reasoning, tool use, and coding; text, image, and video input, 1M-token context."
     },
     capabilities: { vision: true, tools: true, contextWindow: 1e6 }
@@ -1061,6 +1265,7 @@ var VIDEO_MODELS = [
     adapter: "google-genai",
     providerModelId: "veo-3.1-generate-preview",
     lifecycle: "active",
+    exposure: "catalog",
     constraints: {
       aspectRatios: [...OMNI_RATIOS],
       resolutions: ["720p", "1080p", "4k"],
@@ -1069,7 +1274,7 @@ var VIDEO_MODELS = [
       supports: ["lastFrame"]
     },
     defaults: { durationSeconds: 8, resolution: "720p", aspectRatio: "16:9" },
-    usage: { ...VIDEO_USAGE },
+    usage: { ...VIDEO_USAGE, slug: "veo" },
     pricingKind: "video_per_second",
     display: {
       provider: "google",
@@ -1090,6 +1295,7 @@ var VIDEO_MODELS = [
     adapter: "google-genai",
     providerModelId: "veo-3.1-fast-generate-preview",
     lifecycle: "active",
+    exposure: "internal",
     constraints: {
       aspectRatios: [...OMNI_RATIOS],
       resolutions: ["720p", "1080p", "4k"],
@@ -1118,9 +1324,10 @@ var VIDEO_MODELS = [
     adapter: "google-genai",
     providerModelId: "veo-3.1-lite-generate-preview",
     lifecycle: "active",
+    exposure: "internal",
     constraints: {
       aspectRatios: [...OMNI_RATIOS],
-      resolutions: ["720p", "1080p", "4k"],
+      resolutions: ["720p", "1080p"],
       durations: [4, 8],
       references: { images: 3 },
       supports: ["lastFrame"]
@@ -1148,6 +1355,7 @@ var VIDEO_MODELS = [
     adapter: "google-genai",
     providerModelId: "gemini-omni-flash-preview",
     lifecycle: "active",
+    exposure: "catalog",
     constraints: {
       // 1.0 rejects `resolution`; it renders 720p. No defaults: the adapter
       // sends ratio and duration only when the caller set them, so a
@@ -1158,7 +1366,7 @@ var VIDEO_MODELS = [
       supports: []
     },
     defaults: {},
-    usage: { ...VIDEO_USAGE },
+    usage: { ...VIDEO_USAGE, slug: "gemini_omni" },
     pricingKind: "video_per_second",
     display: {
       provider: "google",
@@ -1175,6 +1383,7 @@ var VIDEO_MODELS = [
     adapter: "google-genai",
     providerModelId: "gemini-omni-1.1-flash",
     lifecycle: "active",
+    exposure: "catalog",
     constraints: {
       aspectRatios: [...OMNI_RATIOS],
       resolutions: ["360p", "720p", "1080p", "4k"],
@@ -1200,10 +1409,12 @@ var VIDEO_MODELS = [
     adapter: "ark",
     providerModelId: "dreamina-seedance-2-0-260128",
     lifecycle: "active",
+    exposure: "catalog",
     constraints: {
       aspectRatios: [...SEEDANCE_RATIOS],
       resolutions: ["480p", "720p", "1080p", "4k"],
       durations: [4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15],
+      durationPresets: [4, 6, 8, 10, 12, 15],
       references: { images: 9, videos: 3, audios: 3, audioOnly: false },
       ratioIgnoredWith: ["firstFrame"],
       supports: ["generateAudio", "seed", "lastFrame", "watermark"]
@@ -1214,7 +1425,7 @@ var VIDEO_MODELS = [
       aspectRatio: "16:9",
       generateAudio: true
     },
-    usage: { ...VIDEO_USAGE },
+    usage: { ...VIDEO_USAGE, slug: "seedance2", providerLabel: "byteplus" },
     pricingKind: "video_per_second",
     display: {
       provider: "bytedance",
@@ -1231,6 +1442,7 @@ var VIDEO_MODELS = [
     adapter: "ark",
     providerModelId: "dreamina-seedance-2-0-fast-260128",
     lifecycle: "active",
+    exposure: "catalog",
     constraints: {
       aspectRatios: [...SEEDANCE_RATIOS],
       resolutions: ["480p", "720p"],
@@ -1262,6 +1474,7 @@ var VIDEO_MODELS = [
     adapter: "ark",
     providerModelId: "dreamina-seedance-2-5-260628",
     lifecycle: "active",
+    exposure: "catalog",
     constraints: {
       aspectRatios: [...SEEDANCE_RATIOS],
       resolutions: ["480p", "720p", "1080p"],
@@ -1293,10 +1506,12 @@ var VIDEO_MODELS = [
     adapter: "xai-video",
     providerModelId: "grok-imagine-video",
     lifecycle: "active",
+    exposure: "catalog",
     constraints: {
       aspectRatios: [...GROK_RATIOS],
       resolutions: ["480p", "720p"],
       durations: { min: 1, max: 15 },
+      durationPresets: [4, 6, 8, 10],
       references: { images: 7 },
       supports: []
     },
@@ -1319,6 +1534,7 @@ var VIDEO_MODELS = [
     adapter: "higgsfield",
     providerModelId: "kling-video/v2.5-turbo/pro/image-to-video",
     lifecycle: "active",
+    exposure: "catalog",
     constraints: {
       aspectRatios: [...KLING_RATIOS],
       durations: [5, 10],
@@ -1331,7 +1547,7 @@ var VIDEO_MODELS = [
       cfgScale: 0.5,
       negativePrompt: NEGATIVE_PROMPT
     },
-    usage: { ...VIDEO_USAGE },
+    usage: { ...VIDEO_USAGE, slug: "grok" },
     pricingKind: "video_per_second",
     display: {
       provider: "kuaishou",
@@ -1348,6 +1564,7 @@ var VIDEO_MODELS = [
     adapter: "higgsfield",
     providerModelId: "kling-video/v3.0/pro/image-to-video",
     lifecycle: "active",
+    exposure: "catalog",
     constraints: {
       aspectRatios: [...KLING_RATIOS],
       durations: { min: 3, max: 15 },
@@ -1377,6 +1594,7 @@ var VIDEO_MODELS = [
     adapter: "replicate",
     providerModelId: "kwaivgi/kling-v3-omni-video",
     lifecycle: "active",
+    exposure: "catalog",
     constraints: {
       aspectRatios: [
         "1:1",
@@ -1412,6 +1630,7 @@ var VIDEO_MODELS = [
     adapter: "fal",
     providerModelId: "minimax/h3",
     lifecycle: "active",
+    exposure: "catalog",
     constraints: {
       aspectRatios: [...H3_RATIOS],
       resolutions: ["480p", "768p", "2k", "4k"],
@@ -1440,6 +1659,7 @@ var VIDEO_MODELS = [
     adapter: "fal",
     providerModelId: "minimax/h3-max",
     lifecycle: "active",
+    exposure: "catalog",
     constraints: {
       aspectRatios: [...H3_RATIOS],
       resolutions: ["480p", "768p"],
@@ -1468,6 +1688,7 @@ var VIDEO_MODELS = [
     adapter: "fal",
     providerModelId: "minimax/h3/reference-to-video",
     lifecycle: "active",
+    exposure: "catalog",
     constraints: {
       aspectRatios: [...H3_REFERENCE_RATIOS],
       resolutions: ["480p", "768p", "2k", "4k"],
@@ -1497,6 +1718,7 @@ var VIDEO_MODELS = [
     adapter: "fal",
     providerModelId: "minimax/h3-max/reference-to-video",
     lifecycle: "active",
+    exposure: "catalog",
     constraints: {
       aspectRatios: [...H3_REFERENCE_RATIOS],
       resolutions: ["480p", "768p"],
@@ -1527,6 +1749,7 @@ var VIDEO_MODELS = [
     adapter: "fal",
     providerModelId: "fal-ai/kling-video/o1/video-to-video/reference",
     lifecycle: "active",
+    exposure: "internal",
     constraints: {
       aspectRatios: [...KLING_RATIOS],
       durations: { min: 3, max: 10 },
@@ -1550,6 +1773,7 @@ var VIDEO_MODELS = [
     adapter: "replicate",
     providerModelId: "minimax/hailuo-02",
     lifecycle: "active",
+    exposure: "internal",
     constraints: {
       resolutions: ["768p"],
       durations: [6, 10],
@@ -1572,6 +1796,7 @@ var VIDEO_MODELS = [
     adapter: "higgsfield",
     providerModelId: "dop-preview",
     lifecycle: "active",
+    exposure: "internal",
     constraints: {
       requires: ["firstFrame"],
       supports: ["seed", "lastFrame"]
@@ -1773,6 +1998,40 @@ function resolveVideoOptions(model, request, onUnsupported = "error") {
         model: model.id
       });
   }
+  if (merged.sourceVideo !== void 0 && !model.capabilities?.acceptsSourceVideo) {
+    if (onUnsupported === "error") {
+      throw new ModelKitError({
+        code: "unsupported_option",
+        message: `${model.id} does not take a source video`,
+        model: model.id
+      });
+    }
+    warnings.push({
+      code: "option_ignored",
+      field: "sourceVideo",
+      message: `${model.id} does not take a source video; dropped`
+    });
+    delete merged.sourceVideo;
+  }
+  const hasFrames = merged.firstFrame !== void 0 || merged.lastFrame !== void 0;
+  const hasRefs = compact(merged.referenceImages).length + compact(merged.referenceVideos).length + compact(merged.referenceAudios).length > 0;
+  if (hasFrames && hasRefs && model.capabilities?.framesExcludeReferences) {
+    if (onUnsupported === "error") {
+      throw new ModelKitError({
+        code: "unsupported_option",
+        message: `${model.id} takes frames or references, not both`,
+        model: model.id
+      });
+    }
+    warnings.push({
+      code: "option_ignored",
+      field: "referenceImages",
+      message: `${model.id} takes frames or references, not both; references dropped`
+    });
+    delete merged.referenceImages;
+    delete merged.referenceVideos;
+    delete merged.referenceAudios;
+  }
   for (const input of c.ratioIgnoredWith ?? []) {
     if (merged[input] !== void 0 && merged.aspectRatio !== void 0) {
       if (request.aspectRatio !== void 0) {
@@ -2169,6 +2428,7 @@ function passthroughTextModel(model) {
     adapter: "ai-sdk-text",
     providerModelId: model.modelId,
     lifecycle: "active",
+    exposure: "internal",
     constraints: {},
     defaults: {},
     usage: { unit: "tokens", modality: "TEXT" },