@shop-os/modelkit 0.1.6 → 0.1.7

package/dist/index.js

@@ -83,6 +83,7 @@ var Catalog = class _Catalog {
 
 // src/catalog/image.ts
 var IMAGE_USAGE = { unit: "images", modality: "IMAGE" };
+var IMAGE_GENERATE_EDIT = { imageGenerate: true, imageEdit: true };
 var GEMINI_RATIOS = [
   "1:1",
   "2:3",
@@ -96,7 +97,7 @@ var GEMINI_RATIOS = [
   "5:4",
   "adaptive"
 ];
-function geminiImage(id, label, aliases, sizes) {
+function geminiImage(id, label, aliases, sizes, display) {
   const model = {
     id,
     aliases,
@@ -112,22 +113,40 @@ function geminiImage(id, label, aliases, sizes) {
     },
     defaults: { size: "1K", aspectRatio: "1:1" },
     usage: { ...IMAGE_USAGE },
-    pricingKind: "image_size"
+    pricingKind: "image_size",
+    display,
+    capabilities: { ...IMAGE_GENERATE_EDIT }
   };
   return model;
 }
 var IMAGE_MODELS = [
-  geminiImage("gemini-3-pro-image-preview", "Nano Banana Pro", ["nb-pro"], ["1K", "2K", "4K"]),
+  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-3.1-flash-image-preview",
     "Nano Banana Flash",
     ["nb-2"],
     // Flash 3.1 is the only Gemini image model with the 512 tier.
-    ["512", "1K", "2K", "4K"]
+    ["512", "1K", "2K", "4K"],
+    {
+      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"
+    }
   ),
-  geminiImage("gemini-3.1-flash-lite-image", "Nano Banana Lite", ["nb-lite"], ["1K"]),
+  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"]),
+  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."
+  }),
   {
     // Cowork "auto" lets OpenAI pick the size; the canonical equivalent is adaptive.
     id: "gpt-image-2",
@@ -145,7 +164,13 @@ var IMAGE_MODELS = [
     },
     defaults: { size: "1K", aspectRatio: "adaptive", quality: "auto" },
     usage: { ...IMAGE_USAGE },
-    pricingKind: "image_size"
+    pricingKind: "image_size",
+    display: {
+      provider: "openai",
+      description: "The most capable model for image editing and complex creative instructions. Excels at text rendering, visual consistency, design generation, and iterative refinement.",
+      logo: "https://d31vvfpk2ppdyq.cloudfront.net/assets/image/1204b494-8962-4f7f-af5e-699a8bdc98e3_compressed.png"
+    },
+    capabilities: { ...IMAGE_GENERATE_EDIT }
   },
   {
     // Same Images API surface as gpt-image-2 plus the "xhigh" and "max" quality tiers.
@@ -164,7 +189,12 @@ var IMAGE_MODELS = [
     },
     defaults: { size: "1K", aspectRatio: "adaptive", quality: "auto" },
     usage: { ...IMAGE_USAGE },
-    pricingKind: "image_size"
+    pricingKind: "image_size",
+    display: {
+      provider: "openai",
+      description: "OpenAI's most capable GPT Image 2.5 tier, built for editing precision; takes text and image input and outputs images in low, medium, high, xhigh, or max quality."
+    },
+    capabilities: { ...IMAGE_GENERATE_EDIT }
   },
   {
     // Same Images API surface as gpt-image-2 plus the "xhigh" and "max" quality tiers.
@@ -183,7 +213,12 @@ var IMAGE_MODELS = [
     },
     defaults: { size: "1K", aspectRatio: "adaptive", quality: "auto" },
     usage: { ...IMAGE_USAGE },
-    pricingKind: "image_size"
+    pricingKind: "image_size",
+    display: {
+      provider: "openai",
+      description: "OpenAI's fastest GPT Image 2.5 tier for high-quality everyday image generation and editing; takes text and image input and outputs images."
+    },
+    capabilities: { ...IMAGE_GENERATE_EDIT }
   },
   {
     // Cowork allows 7 ratios; trigger-workflows allows 9 (adds 3:2, 9:21, drops 4:5/5:4).
@@ -215,7 +250,13 @@ var IMAGE_MODELS = [
     },
     defaults: { size: "1K", aspectRatio: "1:1", outputFormat: "png", count: 1 },
     usage: { ...IMAGE_USAGE },
-    pricingKind: "image_size"
+    pricingKind: "image_size",
+    display: {
+      provider: "meta",
+      description: "Meta's image model with built-in world knowledge and strong prompt adherence. Great for concept-rich scenes, realistic compositions, and reference-guided edits.",
+      logo: "https://d31vvfpk2ppdyq.cloudfront.net/assets/characters/1787909065794-k9a7zz.png"
+    },
+    capabilities: { ...IMAGE_GENERATE_EDIT }
   },
   {
     id: "seedream-5-0-pro",
@@ -235,7 +276,12 @@ var IMAGE_MODELS = [
     },
     defaults: { size: "1K", aspectRatio: "1:1" },
     usage: { ...IMAGE_USAGE },
-    pricingKind: "image_size"
+    pricingKind: "image_size",
+    display: {
+      provider: "bytedance",
+      description: "ByteDance Seedream 5.0 Pro text-to-image and multi-reference image-to-image (up to 10 inputs) with region editing; outputs 1K, 1.5K, or 2K."
+    },
+    capabilities: { ...IMAGE_GENERATE_EDIT }
   },
   {
     id: "seedream-4.5",
@@ -267,7 +313,12 @@ var IMAGE_MODELS = [
     },
     defaults: { size: "1K", aspectRatio: "1:1" },
     usage: { ...IMAGE_USAGE },
-    pricingKind: "image_size"
+    pricingKind: "image_size",
+    display: {
+      provider: "bytedance",
+      description: "ByteDance Seedream 4.5 text-to-image and image-to-image with up to 14 reference images and batch image-set generation; outputs 2K or 4K."
+    },
+    capabilities: { ...IMAGE_GENERATE_EDIT }
   },
   {
     id: "flux-2-pro",
@@ -285,7 +336,12 @@ var IMAGE_MODELS = [
     },
     defaults: { size: "1K", aspectRatio: "1:1" },
     usage: { ...IMAGE_USAGE },
-    pricingKind: "image_size"
+    pricingKind: "image_size",
+    display: {
+      provider: "black-forest-labs",
+      description: "Black Forest Labs FLUX.2 [pro] text-to-image and multi-reference editing model served via Prodia; takes a prompt and up to 8 input images and outputs up to 4096x4096."
+    },
+    capabilities: { ...IMAGE_GENERATE_EDIT }
   },
   {
     // Prodia's gemini-3-pro jobs take aspect_ratio + resolution (1K/2K/4K), not
@@ -306,7 +362,12 @@ var IMAGE_MODELS = [
     },
     defaults: { size: "1K", aspectRatio: "1:1" },
     usage: { ...IMAGE_USAGE },
-    pricingKind: "image_size"
+    pricingKind: "image_size",
+    display: {
+      provider: "google",
+      description: "Google Gemini 3 Pro image generation and editing served via Prodia; takes a prompt plus up to 3 reference images and outputs 1K, 2K, or 4K."
+    },
+    capabilities: { ...IMAGE_GENERATE_EDIT }
   },
   {
     // Nano Banana (img2img.v2) is flat-rate with aspect_ratio only; 3:2, 4:5,
@@ -326,7 +387,12 @@ var IMAGE_MODELS = [
     },
     defaults: { size: "1K", aspectRatio: "1:1" },
     usage: { ...IMAGE_USAGE },
-    pricingKind: "image_size"
+    pricingKind: "image_size",
+    display: {
+      provider: "google",
+      description: "Google Nano Banana image generation and editing served via Prodia; takes a prompt plus 1-3 input images and matches the first input's aspect ratio in roughly 8 seconds."
+    },
+    capabilities: { ...IMAGE_GENERATE_EDIT }
   },
   {
     // WAN 2.7 takes `images` (up to 9) and a `size` enum whose exact values
@@ -347,7 +413,12 @@ var IMAGE_MODELS = [
     },
     defaults: { size: "1K", aspectRatio: "1:1" },
     usage: { unit: "compute_seconds", modality: "IMAGE" },
-    pricingKind: "image_size"
+    pricingKind: "image_size",
+    display: {
+      provider: "alibaba",
+      description: "Alibaba Wan 2.7 Image Pro text-to-image and image-editing model; takes a prompt plus up to 9 reference images and generates up to 4K (4096x4096)."
+    },
+    capabilities: { ...IMAGE_GENERATE_EDIT }
   },
   {
     // Qwen Image 2 Pro has aspect_ratio but no size input; resolution follows
@@ -369,7 +440,12 @@ var IMAGE_MODELS = [
     },
     defaults: { aspectRatio: "1:1" },
     usage: { unit: "compute_seconds", modality: "IMAGE" },
-    pricingKind: "image_size"
+    pricingKind: "image_size",
+    display: {
+      provider: "alibaba",
+      description: "Alibaba Qwen Image 2 Pro unified text-to-image and editing model; takes a prompt plus an optional single reference image and generates native 2K output."
+    },
+    capabilities: { ...IMAGE_GENERATE_EDIT }
   },
   {
     // Qwen Image Edit Plus takes aspect_ratio (or match_input_image) and
@@ -391,7 +467,12 @@ var IMAGE_MODELS = [
     },
     defaults: { outputFormat: "png" },
     usage: { unit: "compute_seconds", modality: "IMAGE" },
-    pricingKind: "image_size"
+    pricingKind: "image_size",
+    display: {
+      provider: "alibaba",
+      description: "Alibaba Qwen Image Edit Plus with Lightning and a skin-retouching LoRA; takes one or more reference images plus an edit prompt and outputs webp, jpg, or png."
+    },
+    capabilities: { ...IMAGE_GENERATE_EDIT }
   }
 ];
 
@@ -412,7 +493,11 @@ var OTHER_MODELS = [
     },
     defaults: {},
     usage: { unit: "images", modality: "AUDIO" },
-    pricingKind: "flat"
+    pricingKind: "flat",
+    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."
+    }
   },
   {
     id: "lyria-3-pro-preview",
@@ -428,7 +513,11 @@ var OTHER_MODELS = [
     },
     defaults: {},
     usage: { unit: "images", modality: "AUDIO" },
-    pricingKind: "flat"
+    pricingKind: "flat",
+    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."
+    }
   },
   {
     id: "google/lyria-2",
@@ -444,7 +533,11 @@ var OTHER_MODELS = [
     },
     defaults: {},
     usage: { unit: "compute_seconds", modality: "AUDIO" },
-    pricingKind: "flat"
+    pricingKind: "flat",
+    display: {
+      provider: "google",
+      description: "Lyria 2 music generation on Replicate; produces 30 seconds of 48 kHz stereo audio from a text prompt."
+    }
   },
   {
     id: "mmaudio",
@@ -461,7 +554,11 @@ var OTHER_MODELS = [
     },
     defaults: { durationSeconds: 8 },
     usage: { unit: "compute_seconds", modality: "AUDIO" },
-    pricingKind: "flat"
+    pricingKind: "flat",
+    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."
+    }
   },
   {
     id: "elevenlabs-music",
@@ -478,7 +575,11 @@ var OTHER_MODELS = [
     defaults: { durationSeconds: 15, instrumental: true },
     // Replicate's run() does not report predict_time; billed by generated length.
     usage: { unit: "seconds", modality: "AUDIO" },
-    pricingKind: "flat"
+    pricingKind: "flat",
+    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."
+    }
   },
   {
     id: "topaz",
@@ -495,7 +596,12 @@ var OTHER_MODELS = [
     },
     defaults: { scaleFactor: 2, outputFormat: "png" },
     usage: { unit: "compute_seconds", modality: "IMAGE" },
-    pricingKind: "flat"
+    pricingKind: "flat",
+    display: {
+      provider: "topaz",
+      description: "Topaz Labs AI image upscaling on Replicate; takes one input image and upscales up to 6x with selectable enhancement models and optional face enhancement."
+    },
+    capabilities: { imageEdit: true }
   },
   {
     id: "bg-remove",
@@ -512,7 +618,12 @@ var OTHER_MODELS = [
     },
     defaults: {},
     usage: { unit: "images", modality: "IMAGE" },
-    pricingKind: "flat"
+    pricingKind: "flat",
+    display: {
+      provider: "prodia",
+      description: "Prodia background removal; takes one input image and returns a PNG cut-out of the subject plus an optional black-and-white mask."
+    },
+    capabilities: { imageEdit: true }
   }
 ];
 
@@ -535,22 +646,112 @@ function text(id, label, providerModelId, extra) {
   };
 }
 var TEXT_MODELS = [
-  text("gpt-4o", "GPT-4o", "gpt-4o"),
-  text("gpt-4o-mini", "GPT-4o mini", "gpt-4o-mini"),
-  text("gpt-5.1", "GPT-5.1", "gpt-5.1"),
-  text("gpt-5.2", "GPT-5.2", "gpt-5.2"),
-  text("gpt-5.4", "GPT-5.4", "gpt-5.4"),
+  text("gpt-4o", "GPT-4o", "gpt-4o", {
+    display: {
+      provider: "openai",
+      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", {
+    display: {
+      provider: "openai",
+      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", {
+    display: {
+      provider: "openai",
+      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", {
+    display: {
+      provider: "openai",
+      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", {
+    display: {
+      provider: "openai",
+      description: "OpenAI GPT-5.4 flagship reasoning model; text and image input, 1.05M-token context."
+    },
+    capabilities: {
+      vision: true,
+      tools: true,
+      structuredOutput: true,
+      contextWindow: 105e4
+    }
+  }),
   // 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", {
-    aliases: ["gpt-5.1-mini", "gpt-5.2-mini"]
+    aliases: ["gpt-5.1-mini", "gpt-5.2-mini"],
+    display: {
+      provider: "openai",
+      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", {
+    display: {
+      provider: "openai",
+      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: {
+      vision: true,
+      tools: true,
+      structuredOutput: true,
+      contextWindow: 105e4
+    }
+  }),
+  text("gpt-5.6-luna", "GPT-5.6 Luna", "gpt-5.6-luna", {
+    display: {
+      provider: "openai",
+      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: {
+      vision: true,
+      tools: true,
+      structuredOutput: true,
+      contextWindow: 105e4
+    }
+  }),
+  text("gpt-5.6-terra", "GPT-5.6 Terra", "gpt-5.6-terra", {
+    display: {
+      provider: "openai",
+      description: "OpenAI GPT-5.6 Terra balanced tier of the GPT-5.6 reasoning family; text and image input, 1.05M-token context."
+    },
+    capabilities: {
+      vision: true,
+      tools: true,
+      structuredOutput: true,
+      contextWindow: 105e4
+    }
+  }),
+  text("gpt-5.6-sol", "GPT-5.6 Sol", "gpt-5.6-sol", {
+    display: {
+      provider: "openai",
+      description: "OpenAI GPT-5.6 Sol flagship tier of the GPT-5.6 reasoning family; text and image input, 1.05M-token context."
+    },
+    capabilities: {
+      vision: true,
+      tools: true,
+      structuredOutput: true,
+      contextWindow: 105e4
+    }
   }),
-  text("gpt-5.5", "GPT-5.5", "gpt-5.5"),
-  text("gpt-5.6-luna", "GPT-5.6 Luna", "gpt-5.6-luna"),
-  text("gpt-5.6-terra", "GPT-5.6 Terra", "gpt-5.6-terra"),
-  text("gpt-5.6-sol", "GPT-5.6 Sol", "gpt-5.6-sol"),
   // 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", {
-    aliases: ["claude-sonnet-4.5", "claude-sonnet-4-5-20250929"]
+    aliases: ["claude-sonnet-4.5", "claude-sonnet-4-5-20250929"],
+    display: {
+      provider: "anthropic",
+      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", {
@@ -559,42 +760,280 @@ var TEXT_MODELS = [
       "claude-sonnet-4",
       "claude-sonnet-4.0",
       "claude-sonnet-4-20250514"
-    ]
+    ],
+    display: {
+      provider: "anthropic",
+      description: "Anthropic Claude Sonnet 4.6 (legacy) balanced model; text and image input, 1M-token context."
+    },
+    capabilities: {
+      vision: true,
+      tools: true,
+      structuredOutput: true,
+      contextWindow: 1e6
+    }
+  }),
+  text("claude-sonnet-5", "Claude Sonnet 5", "claude-sonnet-5", {
+    display: {
+      provider: "anthropic",
+      description: "Anthropic Claude Sonnet 5 current balanced model; text and image input, 1M-token context."
+    },
+    capabilities: {
+      vision: true,
+      tools: true,
+      structuredOutput: true,
+      contextWindow: 1e6
+    }
   }),
-  text("claude-sonnet-5", "Claude Sonnet 5", "claude-sonnet-5"),
   text("claude-opus-4-8", "Claude Opus 4.8", "claude-opus-4-8", {
-    aliases: ["claude-opus-4.8", "claude-opus-4", "claude-opus-4.0", "claude-opus-4-20250514"]
+    aliases: ["claude-opus-4.8", "claude-opus-4", "claude-opus-4.0", "claude-opus-4-20250514"],
+    display: {
+      provider: "anthropic",
+      description: "Anthropic Claude Opus 4.8 (legacy) frontier model; text and image input, 1M-token context."
+    },
+    capabilities: {
+      vision: true,
+      tools: true,
+      structuredOutput: true,
+      contextWindow: 1e6
+    }
   }),
   text("claude-haiku-4-5", "Claude Haiku 4.5", "claude-haiku-4-5", {
-    aliases: ["claude-haiku-4.5", "claude-haiku-4-5-20251001", "claude-haiku-4"]
+    aliases: ["claude-haiku-4.5", "claude-haiku-4-5-20251001", "claude-haiku-4"],
+    display: {
+      provider: "anthropic",
+      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 }
   }),
   text("claude-opus-5", "Claude Opus 5", "claude-opus-5", {
-    aliases: ["claude-opus-5.0"]
+    aliases: ["claude-opus-5.0"],
+    display: {
+      provider: "anthropic",
+      description: "Anthropic Claude Opus 5 current frontier model for agentic coding and enterprise work; text and image input, 1M-token context."
+    },
+    capabilities: {
+      vision: true,
+      tools: true,
+      structuredOutput: true,
+      contextWindow: 1e6
+    }
+  }),
+  text("claude-fable-5", "Claude Fable 5", "claude-fable-5", {
+    display: {
+      provider: "anthropic",
+      description: "Anthropic Claude Fable 5 (legacy) model for demanding reasoning and long-horizon agentic work; text and image input, 1M-token context."
+    },
+    capabilities: {
+      vision: true,
+      tools: true,
+      structuredOutput: true,
+      contextWindow: 1e6
+    }
+  }),
+  text("gemini-2.5-flash", "Gemini 2.5 Flash", "gemini-2.5-flash", {
+    display: {
+      provider: "google",
+      description: "Google Gemini 2.5 Flash low-latency multimodal model for high-volume reasoning over text, images, video, and audio; 1M-token context."
+    },
+    capabilities: {
+      vision: true,
+      tools: true,
+      structuredOutput: true,
+      contextWindow: 1048576
+    }
+  }),
+  text("gemini-2.5-pro", "Gemini 2.5 Pro", "gemini-2.5-pro", {
+    display: {
+      provider: "google",
+      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: {
+      vision: true,
+      tools: true,
+      structuredOutput: true,
+      contextWindow: 1048576
+    }
+  }),
+  text("gemini-3.1-pro-preview", "Gemini 3.1 Pro (preview)", "gemini-3.1-pro-preview", {
+    display: {
+      provider: "google",
+      description: "Google Gemini 3.1 Pro preview reasoning model for software engineering, tool use, and reliable agentic workflows; 1M-token context."
+    },
+    capabilities: {
+      vision: true,
+      tools: true,
+      structuredOutput: true,
+      contextWindow: 1048576
+    }
+  }),
+  text("gemini-3.5-flash", "Gemini 3.5 Flash", "gemini-3.5-flash", {
+    display: {
+      provider: "google",
+      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: {
+      vision: true,
+      tools: true,
+      structuredOutput: true,
+      contextWindow: 1048576
+    }
+  }),
+  text("gemini-3.6-flash", "Gemini 3.6 Flash", "gemini-3.6-flash", {
+    display: {
+      provider: "google",
+      description: "Google Gemini 3.6 Flash general-purpose multimodal agentic model balancing speed and capability; 1M-token context."
+    },
+    capabilities: {
+      vision: true,
+      tools: true,
+      structuredOutput: true,
+      contextWindow: 1048576
+    }
+  }),
+  text("gemini-3.7-flash", "Gemini 3.7 Flash", "gemini-3.7-flash", {
+    display: {
+      provider: "google",
+      description: "Google Gemini 3.7 Flash natively multimodal reasoning model for coding, agentic workflows, and multi-step execution; 1M-token context."
+    },
+    capabilities: {
+      vision: true,
+      tools: true,
+      structuredOutput: true,
+      contextWindow: 1048576
+    }
+  }),
+  text("gemini-3.8-flash", "Gemini 3.8 Flash", "gemini-3.8-flash", {
+    display: {
+      provider: "google",
+      description: "Google Gemini 3.8 Flash frontier model for long-horizon software engineering, autonomous agents, and enterprise workflows; 1M-token context."
+    },
+    capabilities: {
+      vision: true,
+      tools: true,
+      structuredOutput: true,
+      contextWindow: 1048576
+    }
   }),
-  text("claude-fable-5", "Claude Fable 5", "claude-fable-5"),
-  text("gemini-2.5-flash", "Gemini 2.5 Flash", "gemini-2.5-flash"),
-  text("gemini-2.5-pro", "Gemini 2.5 Pro", "gemini-2.5-pro"),
-  text("gemini-3.1-pro-preview", "Gemini 3.1 Pro (preview)", "gemini-3.1-pro-preview"),
-  text("gemini-3.5-flash", "Gemini 3.5 Flash", "gemini-3.5-flash"),
-  text("gemini-3.6-flash", "Gemini 3.6 Flash", "gemini-3.6-flash"),
-  text("gemini-3.7-flash", "Gemini 3.7 Flash", "gemini-3.7-flash"),
-  text("gemini-3.8-flash", "Gemini 3.8 Flash", "gemini-3.8-flash"),
   // 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", {
     lifecycle: "deprecated",
     replacedBy: "gemini-3.6-flash"
   }),
-  text("gemini-3.1-flash-lite", "Gemini 3.1 Flash Lite", "gemini-3.1-flash-lite"),
-  text("sonar", "Sonar", "sonar"),
-  text("sonar-pro", "Sonar Pro", "sonar-pro"),
-  text("sonar-reasoning-pro", "Sonar Reasoning Pro", "sonar-reasoning-pro"),
-  text("deepseek-v4-flash", "DeepSeek V4 Flash", "deepseek-v4-flash"),
-  text("deepseek-v4-pro", "DeepSeek V4 Pro", "deepseek-v4-pro"),
-  text("grok-4.3", "Grok 4.3", "grok-4.3"),
-  text("kimi-k2.5", "Kimi K2.5", "kimi-k2.5"),
-  text("kimi-k3", "Kimi K3", "kimi-k3"),
+  text("gemini-3.1-flash-lite", "Gemini 3.1 Flash Lite", "gemini-3.1-flash-lite", {
+    display: {
+      provider: "google",
+      description: "Google Gemini 3.1 Flash-Lite low-latency multimodal model for high-volume agentic, translation, and data-extraction tasks; 1M-token context."
+    },
+    capabilities: {
+      vision: true,
+      tools: true,
+      structuredOutput: true,
+      contextWindow: 1048576
+    }
+  }),
+  text("sonar", "Sonar", "sonar", {
+    display: {
+      provider: "perplexity",
+      description: "Perplexity Sonar lightweight web-grounded search model returning cited answers; 128K-token context."
+    },
+    capabilities: {
+      vision: true,
+      tools: false,
+      structuredOutput: true,
+      contextWindow: 128e3
+    }
+  }),
+  text("sonar-pro", "Sonar Pro", "sonar-pro", {
+    display: {
+      provider: "perplexity",
+      description: "Perplexity Sonar Pro advanced web-grounded search model for complex multi-step queries; 200K-token context."
+    },
+    capabilities: {
+      vision: true,
+      tools: false,
+      structuredOutput: true,
+      contextWindow: 2e5
+    }
+  }),
+  text("sonar-reasoning-pro", "Sonar Reasoning Pro", "sonar-reasoning-pro", {
+    display: {
+      provider: "perplexity",
+      description: "Perplexity Sonar Reasoning Pro chain-of-thought web-grounded search model for complex problem-solving; 128K-token context."
+    },
+    capabilities: {
+      vision: true,
+      tools: false,
+      structuredOutput: true,
+      contextWindow: 128e3
+    }
+  }),
+  text("deepseek-v4-flash", "DeepSeek V4 Flash", "deepseek-v4-flash", {
+    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."
+    },
+    capabilities: {
+      vision: true,
+      tools: true,
+      structuredOutput: true,
+      contextWindow: 1e6
+    }
+  }),
+  text("deepseek-v4-pro", "DeepSeek V4 Pro", "deepseek-v4-pro", {
+    display: {
+      provider: "deepseek",
+      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: {
+      vision: false,
+      tools: true,
+      structuredOutput: true,
+      contextWindow: 1e6
+    }
+  }),
+  // 1M-token context: https://docs.x.ai/docs/models/grok-4.3
+  text("grok-4.3", "Grok 4.3", "grok-4.3", {
+    display: {
+      provider: "xai",
+      description: "xAI Grok 4.3 reasoning model with tool calling and structured outputs; text and image input, 1M-token context."
+    },
+    capabilities: {
+      vision: true,
+      tools: true,
+      structuredOutput: true,
+      contextWindow: 1e6
+    }
+  }),
+  // 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", {
+    display: {
+      provider: "moonshot",
+      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", {
+    display: {
+      provider: "moonshot",
+      description: "Moonshot Kimi K3 flagship model with native visual understanding and always-on thinking; text, image, and video input, 1M-token context."
+    },
+    capabilities: {
+      vision: true,
+      tools: true,
+      structuredOutput: true,
+      contextWindow: 1e6
+    }
+  }),
   // Cowork calls the provider "MiniMax-M3"; the pricing key is lowercase.
-  text("minimax-m3", "MiniMax M3", "MiniMax-M3", { aliases: ["MiniMax-M3"] })
+  // 1M-token context: https://platform.minimax.io/docs/api-reference/text-openai-api
+  text("minimax-m3", "MiniMax M3", "MiniMax-M3", {
+    aliases: ["MiniMax-M3"],
+    display: {
+      provider: "minimax",
+      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 }
+  })
 ];
 
 // src/catalog/video.ts
@@ -606,6 +1045,8 @@ var GROK_RATIOS = ["1:1", "16:9", "9:16", "4:3", "3:4", "3:2", "2:3"];
 var H3_RATIOS = ["21:9", "16:9", "4:3", "1:1", "3:4", "9:16"];
 var H3_REFERENCE_RATIOS = ["adaptive", ...H3_RATIOS];
 var NEGATIVE_PROMPT = "blur, distort, and low quality";
+var FRAMES_EXCLUDE_REFERENCES = { framesExcludeReferences: true };
+var GOOGLE_LOGO = "https://d31vvfpk2ppdyq.cloudfront.net/assets/image/192485d2-a76a-4657-9c06-930b2c70badd_compressed.png";
 var VIDEO_MODELS = [
   {
     id: "veo-3.1",
@@ -629,7 +1070,13 @@ var VIDEO_MODELS = [
     },
     defaults: { durationSeconds: 8, resolution: "720p", aspectRatio: "16:9" },
     usage: { ...VIDEO_USAGE },
-    pricingKind: "video_per_second"
+    pricingKind: "video_per_second",
+    display: {
+      provider: "google",
+      description: "Google's highest-quality video generation model. Supports up to 4K resolution and 8-second clips with reference image conditioning.",
+      logo: GOOGLE_LOGO
+    },
+    capabilities: { ...FRAMES_EXCLUDE_REFERENCES }
   },
   {
     id: "veo-3.1-fast",
@@ -652,7 +1099,12 @@ var VIDEO_MODELS = [
     },
     defaults: { durationSeconds: 8, resolution: "720p", aspectRatio: "16:9" },
     usage: { ...VIDEO_USAGE },
-    pricingKind: "video_per_second"
+    pricingKind: "video_per_second",
+    display: {
+      provider: "google",
+      description: "Faster variant of Google Veo 3.1; text- and image-to-video with natively generated audio at 4 or 8 seconds."
+    },
+    capabilities: { ...FRAMES_EXCLUDE_REFERENCES }
   },
   {
     id: "veo-3.1-lite",
@@ -675,7 +1127,12 @@ var VIDEO_MODELS = [
     },
     defaults: { durationSeconds: 8, resolution: "720p", aspectRatio: "16:9" },
     usage: { ...VIDEO_USAGE },
-    pricingKind: "video_per_second"
+    pricingKind: "video_per_second",
+    display: {
+      provider: "google",
+      description: "Efficiency-focused Google Veo 3.1 Lite for high-volume video generation and editing; text and image input, with no 4K output or video extension."
+    },
+    capabilities: { ...FRAMES_EXCLUDE_REFERENCES }
   },
   {
     id: "gemini-omni-flash",
@@ -702,7 +1159,13 @@ var VIDEO_MODELS = [
     },
     defaults: {},
     usage: { ...VIDEO_USAGE },
-    pricingKind: "video_per_second"
+    pricingKind: "video_per_second",
+    display: {
+      provider: "google",
+      description: "Google's conversational video model with native audio, human support, and natural language editing. Aspect ratio only \u2014 resolution and duration are model-determined.",
+      logo: GOOGLE_LOGO
+    },
+    capabilities: { acceptsSourceVideo: true }
   },
   {
     id: "gemini-omni-1.1-flash",
@@ -721,7 +1184,12 @@ var VIDEO_MODELS = [
     },
     defaults: {},
     usage: { ...VIDEO_USAGE },
-    pricingKind: "video_per_second"
+    pricingKind: "video_per_second",
+    display: {
+      provider: "google",
+      description: "Google Gemini Omni Flash 1.1 conversational video generation and editing; text, image, and up to 10-second video input, 3-10 second output at up to 4K and 24 FPS."
+    },
+    capabilities: { acceptsSourceVideo: true }
   },
   {
     // vibe-triggers and Space Builder drop the ratio when a frame is wired; Cowork always sends it.
@@ -747,7 +1215,13 @@ var VIDEO_MODELS = [
       generateAudio: true
     },
     usage: { ...VIDEO_USAGE },
-    pricingKind: "video_per_second"
+    pricingKind: "video_per_second",
+    display: {
+      provider: "bytedance",
+      description: "High-quality AI video generation with reference image conditioning. Supports 4-15 second clips with human preprocessing.",
+      logo: "https://d31vvfpk2ppdyq.cloudfront.net/assets/image/3bba5d81-4e57-4e71-911b-cdfa05acca2c_compressed.png"
+    },
+    capabilities: { acceptsSourceVideo: true, ...FRAMES_EXCLUDE_REFERENCES }
   },
   {
     id: "seedance-2.0-fast",
@@ -772,7 +1246,12 @@ var VIDEO_MODELS = [
       generateAudio: true
     },
     usage: { ...VIDEO_USAGE },
-    pricingKind: "video_per_second"
+    pricingKind: "video_per_second",
+    display: {
+      provider: "bytedance",
+      description: "Speed- and cost-optimized ByteDance Seedance 2.0 Fast; text plus image, video, and audio references to generate or edit 4-15 second clips."
+    },
+    capabilities: { acceptsSourceVideo: true, ...FRAMES_EXCLUDE_REFERENCES }
   },
   {
     // generateAudio defaults to false here (trigger-workflows); Space Builder defaults it to true.
@@ -798,7 +1277,12 @@ var VIDEO_MODELS = [
       generateAudio: false
     },
     usage: { ...VIDEO_USAGE },
-    pricingKind: "video_per_second"
+    pricingKind: "video_per_second",
+    display: {
+      provider: "bytedance",
+      description: "ByteDance Seedance 2.5 for long-form storytelling and omni-reference generation/editing; up to 50 reference assets and 4-30 second clips at up to 1080p."
+    },
+    capabilities: { acceptsSourceVideo: true, ...FRAMES_EXCLUDE_REFERENCES }
   },
   {
     // Cowork allows 7 reference images and durations 4/6/8/10; Space Builder allows 5 and 1-15s.
@@ -819,7 +1303,13 @@ var VIDEO_MODELS = [
     // No resolution default: xAI picks its own unless the caller asks.
     defaults: { durationSeconds: 8, aspectRatio: "16:9" },
     usage: { unit: "reported_cost_usd", modality: "VIDEO" },
-    pricingKind: "video_per_second"
+    pricingKind: "video_per_second",
+    display: {
+      provider: "xai",
+      description: "xAI's reference-to-video model. Generates short clips conditioned on up to 7 reference images with flexible aspect ratios.",
+      logo: "https://d31vvfpk2ppdyq.cloudfront.net/assets/image/858b7fcc-9bd5-411d-aea1-a7b35ecfffab_compressed.png"
+    },
+    capabilities: { ...FRAMES_EXCLUDE_REFERENCES }
   },
   {
     id: "kling-2.5",
@@ -842,7 +1332,12 @@ var VIDEO_MODELS = [
       negativePrompt: NEGATIVE_PROMPT
     },
     usage: { ...VIDEO_USAGE },
-    pricingKind: "video_per_second"
+    pricingKind: "video_per_second",
+    display: {
+      provider: "kuaishou",
+      description: "Kuaishou Kling 2.5 Turbo Pro image-to-video; animates a single input image from a text prompt with an optional end image, 5- or 10-second clips."
+    },
+    capabilities: { ...FRAMES_EXCLUDE_REFERENCES }
   },
   {
     // trigger-workflows validates 5/10 only; Space Builder allows 3-15s (Kling30Input).
@@ -866,7 +1361,12 @@ var VIDEO_MODELS = [
       negativePrompt: NEGATIVE_PROMPT
     },
     usage: { ...VIDEO_USAGE },
-    pricingKind: "video_per_second"
+    pricingKind: "video_per_second",
+    display: {
+      provider: "kuaishou",
+      description: "Kuaishou Kling 3.0 Pro image-to-video; a required start image plus a text prompt (or multi-shot prompts) with optional end image and native audio, 3-15 second clips."
+    },
+    capabilities: { ...FRAMES_EXCLUDE_REFERENCES }
   },
   {
     // mode standard/pro selects 720p/1080p; the endpoint takes no separate resolution.
@@ -897,7 +1397,12 @@ var VIDEO_MODELS = [
     },
     defaults: { durationSeconds: 5, aspectRatio: "16:9", generateAudio: false, mode: "pro" },
     usage: { unit: "compute_seconds", modality: "VIDEO" },
-    pricingKind: "video_per_second"
+    pricingKind: "video_per_second",
+    display: {
+      provider: "kuaishou",
+      description: "Kuaishou Kling Video 3.0 Omni unified generation and editing from text, start/end images, up to 7 reference images, and reference video; up to 15 seconds at up to 1080p."
+    },
+    capabilities: {}
   },
   {
     id: "minimax-h3",
@@ -920,7 +1425,12 @@ var VIDEO_MODELS = [
       promptExpansion: "balanced"
     },
     usage: { ...VIDEO_USAGE },
-    pricingKind: "video_per_second"
+    pricingKind: "video_per_second",
+    display: {
+      provider: "minimax",
+      description: "MiniMax H3 multimodal video model for text-to-video, image-to-video (first/last frame), and reference-to-video with native audio, up to 2K."
+    },
+    capabilities: { ...FRAMES_EXCLUDE_REFERENCES }
   },
   {
     id: "minimax-h3-max",
@@ -943,7 +1453,12 @@ var VIDEO_MODELS = [
       promptExpansion: "balanced"
     },
     usage: { ...VIDEO_USAGE },
-    pricingKind: "video_per_second"
+    pricingKind: "video_per_second",
+    display: {
+      provider: "minimax",
+      description: "High-speed MiniMax H3 Max variant for text-to-video and image-to-video (first/last frame) only, up to 768p."
+    },
+    capabilities: { ...FRAMES_EXCLUDE_REFERENCES }
   },
   {
     id: "minimax-h3-reference",
@@ -967,7 +1482,12 @@ var VIDEO_MODELS = [
       promptExpansion: "balanced"
     },
     usage: { ...VIDEO_USAGE },
-    pricingKind: "video_per_second"
+    pricingKind: "video_per_second",
+    display: {
+      provider: "minimax",
+      description: "MiniMax H3 reference-to-video mode guided by up to 9 reference images, 3 reference videos, and 3 audio clips; reference and image-to-video inputs cannot be mixed."
+    },
+    capabilities: {}
   },
   {
     id: "minimax-h3-max-reference",
@@ -991,7 +1511,12 @@ var VIDEO_MODELS = [
       promptExpansion: "balanced"
     },
     usage: { ...VIDEO_USAGE },
-    pricingKind: "video_per_second"
+    pricingKind: "video_per_second",
+    display: {
+      provider: "minimax",
+      description: "MiniMax H3 Max reference-to-video on fal; conditions on reference images, videos, and audio. MiniMax's own docs do not list a reference mode for H3 Max."
+    },
+    capabilities: {}
   },
   {
     // Duration 3-10s per trigger-workflows mapDuration; the endpoint takes no resolution.
@@ -1010,7 +1535,12 @@ var VIDEO_MODELS = [
     },
     defaults: { durationSeconds: 5, aspectRatio: "16:9" },
     usage: { ...VIDEO_USAGE },
-    pricingKind: "video_per_second"
+    pricingKind: "video_per_second",
+    display: {
+      provider: "kuaishou",
+      description: "Kuaishou Kling O1 Reference Video-to-Video; a prompt plus a 3-10 second reference video (optional reference images) generates 3-10 second clips preserving motion and camera language."
+    },
+    capabilities: { acceptsSourceVideo: true }
   },
   {
     id: "hailuo-02",
@@ -1027,7 +1557,12 @@ var VIDEO_MODELS = [
     },
     defaults: { durationSeconds: 6, resolution: "768p" },
     usage: { unit: "compute_seconds", modality: "VIDEO" },
-    pricingKind: "video_per_second"
+    pricingKind: "video_per_second",
+    display: {
+      provider: "minimax",
+      description: "MiniMax Hailuo 02 text-to-video and image-to-video model with realistic physics and strong prompt following; outputs 768p at 6 or 10 seconds."
+    },
+    capabilities: {}
   },
   {
     id: "higgsfield-dop-preview",
@@ -1043,7 +1578,12 @@ var VIDEO_MODELS = [
     },
     defaults: {},
     usage: { ...VIDEO_USAGE },
-    pricingKind: "video_per_second"
+    pricingKind: "video_per_second",
+    display: {
+      provider: "higgsfield",
+      description: "Higgsfield DOP image-to-video model; animates a required first-frame image from a text prompt with an optional last-frame image and motion controls."
+    },
+    capabilities: {}
   }
 ];