@shop-os/modelkit 0.1.10 → 0.1.11

package/dist/index.js

@@ -99,8 +99,7 @@ var Catalog = class _Catalog {
 };
 
 // src/catalog/image.ts
-var IMAGE_USAGE = { unit: "images", modality: "IMAGE" };
-var IMAGE_GENERATE_EDIT = { imageGenerate: true, imageEdit: true };
+var IMAGE_USAGE = { unit: "images" };
 var GEMINI_RATIOS = [
   "1:1",
   "2:3",
@@ -132,8 +131,7 @@ function geminiImage(id, label, aliases, sizes, display, exposure) {
     defaults: { size: "1K", aspectRatio: "1:1" },
     usage: { ...IMAGE_USAGE, slug: "gemini" },
     pricingKind: "image_size",
-    display,
-    capabilities: { ...IMAGE_GENERATE_EDIT }
+    display
   };
   return model;
 }
@@ -220,8 +218,7 @@ var IMAGE_MODELS = [
       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.
@@ -255,8 +252,7 @@ var IMAGE_MODELS = [
     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.
@@ -290,8 +286,7 @@ var IMAGE_MODELS = [
     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).
@@ -329,8 +324,7 @@ var IMAGE_MODELS = [
       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",
@@ -356,8 +350,7 @@ var IMAGE_MODELS = [
       provider: "bytedance",
       logo: "https://d31vvfpk2ppdyq.cloudfront.net/assets/image/3bba5d81-4e57-4e71-911b-cdfa05acca2c_compressed.png",
       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",
@@ -394,8 +387,7 @@ var IMAGE_MODELS = [
     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",
@@ -417,9 +409,9 @@ var IMAGE_MODELS = [
     pricingKind: "image_size",
     display: {
       provider: "black-forest-labs",
+      logo: "https://cdn.jsdelivr.net/npm/@lobehub/icons-static-png@1.97.0/light/flux.png",
       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
@@ -445,8 +437,7 @@ var IMAGE_MODELS = [
     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,
@@ -471,8 +462,7 @@ var IMAGE_MODELS = [
     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
@@ -493,13 +483,12 @@ var IMAGE_MODELS = [
       supports: []
     },
     defaults: { size: "1K", aspectRatio: "1:1" },
-    usage: { unit: "compute_seconds", modality: "IMAGE" },
+    usage: { unit: "compute_seconds" },
     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
@@ -521,13 +510,12 @@ var IMAGE_MODELS = [
       supports: []
     },
     defaults: { aspectRatio: "1:1" },
-    usage: { unit: "compute_seconds", modality: "IMAGE" },
+    usage: { unit: "compute_seconds" },
     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
@@ -549,13 +537,12 @@ var IMAGE_MODELS = [
       supports: ["outputFormat"]
     },
     defaults: { outputFormat: "png" },
-    usage: { unit: "compute_seconds", modality: "IMAGE" },
+    usage: { unit: "compute_seconds" },
     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 }
+    }
   }
 ];
 
@@ -577,7 +564,7 @@ var OTHER_MODELS = [
       supports: []
     },
     defaults: {},
-    usage: { unit: "images", modality: "AUDIO" },
+    usage: { unit: "images" },
     pricingKind: "flat",
     display: {
       provider: "google",
@@ -600,7 +587,7 @@ var OTHER_MODELS = [
       supports: []
     },
     defaults: {},
-    usage: { unit: "images", modality: "AUDIO" },
+    usage: { unit: "images" },
     pricingKind: "flat",
     display: {
       provider: "google",
@@ -623,7 +610,7 @@ var OTHER_MODELS = [
       supports: ["seed", "negativePrompt"]
     },
     defaults: {},
-    usage: { unit: "compute_seconds", modality: "AUDIO" },
+    usage: { unit: "compute_seconds" },
     pricingKind: "flat",
     display: {
       provider: "google",
@@ -646,7 +633,7 @@ var OTHER_MODELS = [
       supports: ["seed", "negativePrompt"]
     },
     defaults: { durationSeconds: 8 },
-    usage: { unit: "compute_seconds", modality: "AUDIO" },
+    usage: { unit: "compute_seconds" },
     pricingKind: "flat",
     display: {
       provider: "mmaudio",
@@ -669,7 +656,7 @@ 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" },
+    usage: { unit: "seconds" },
     pricingKind: "flat",
     display: {
       provider: "elevenlabs",
@@ -692,13 +679,13 @@ var OTHER_MODELS = [
       supports: ["scaleFactor", "outputFormat", "upscale"]
     },
     defaults: { scaleFactor: 2, outputFormat: "png" },
-    usage: { unit: "compute_seconds", modality: "IMAGE" },
+    usage: { unit: "compute_seconds" },
     pricingKind: "flat",
     display: {
       provider: "topaz",
+      logo: "https://cdn.jsdelivr.net/npm/@lobehub/icons-static-png@1.97.0/light/topazlabs.png",
       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",
@@ -715,19 +702,30 @@ var OTHER_MODELS = [
       supports: []
     },
     defaults: {},
-    usage: { unit: "images", modality: "IMAGE" },
+    usage: { unit: "images" },
     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 }
+    }
   }
 ];
 
 // src/catalog/text.ts
-var TEXT_USAGE = { unit: "tokens", modality: "TEXT" };
+var TEXT_USAGE = { unit: "tokens" };
+var CDN = "https://d31vvfpk2ppdyq.cloudfront.net/assets/image";
+var PROVIDER_LOGOS = {
+  anthropic: `${CDN}/25fad392-188c-4829-a3d0-c4fa02d1c8cd.png`,
+  openai: `${CDN}/1204b494-8962-4f7f-af5e-699a8bdc98e3.png`,
+  google: `${CDN}/192485d2-a76a-4657-9c06-930b2c70badd.png`,
+  deepseek: `${CDN}/9b7406d2-a696-4558-ab79-90a37e3f4574.png`,
+  xai: `${CDN}/858b7fcc-9bd5-411d-aea1-a7b35ecfffab.png`,
+  moonshot: `${CDN}/5a1e2eae-3474-4fac-9c74-a93f41ee6f67.png`,
+  minimax: `${CDN}/f4c2e906-28b7-4a4a-b491-df8fef7808ed.png`
+};
 function text(id, label, providerModelId, extra) {
+  const display = extra?.display;
+  const logo = display && !display.logo ? PROVIDER_LOGOS[display.provider] : void 0;
   return {
     id,
     aliases: [],
@@ -741,7 +739,8 @@ function text(id, label, providerModelId, extra) {
     defaults: {},
     usage: { ...TEXT_USAGE },
     pricingKind: "token",
-    ...extra
+    ...extra,
+    ...display && logo ? { display: { ...display, logo } } : {}
   };
 }
 var ANTHROPIC_THINKING = {
@@ -760,7 +759,7 @@ var TEXT_MODELS = [
       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 }
+    capabilities: { vision: true, contextWindow: 128e3 }
   }),
   text("gpt-4o-mini", "GPT-4o mini", "gpt-4o-mini", {
     exposure: "internal",
@@ -769,7 +768,7 @@ var TEXT_MODELS = [
       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 }
+    capabilities: { vision: true, contextWindow: 128e3 }
   }),
   text("gpt-5.1", "GPT-5.1", "gpt-5.1", {
     exposure: "catalog",
@@ -778,7 +777,7 @@ var TEXT_MODELS = [
       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 }
+    capabilities: { vision: true, contextWindow: 4e5 }
   }),
   text("gpt-5.2", "GPT-5.2", "gpt-5.2", {
     exposure: "catalog",
@@ -787,7 +786,7 @@ var TEXT_MODELS = [
       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 }
+    capabilities: { vision: 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", {
@@ -799,8 +798,6 @@ var TEXT_MODELS = [
     },
     capabilities: {
       vision: true,
-      tools: true,
-      structuredOutput: true,
       contextWindow: 105e4
     }
   }),
@@ -813,7 +810,7 @@ var TEXT_MODELS = [
       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 }
+    capabilities: { vision: true, contextWindow: 4e5 }
   }),
   text("gpt-5.5", "GPT-5.5", "gpt-5.5", {
     exposure: "catalog",
@@ -824,8 +821,6 @@ var TEXT_MODELS = [
     },
     capabilities: {
       vision: true,
-      tools: true,
-      structuredOutput: true,
       contextWindow: 105e4
     }
   }),
@@ -840,8 +835,6 @@ var TEXT_MODELS = [
     capabilities: {
       reasoning: true,
       vision: true,
-      tools: true,
-      structuredOutput: true,
       contextWindow: 105e4
     }
   }),
@@ -856,8 +849,6 @@ var TEXT_MODELS = [
     capabilities: {
       reasoning: true,
       vision: true,
-      tools: true,
-      structuredOutput: true,
       contextWindow: 105e4
     }
   }),
@@ -872,8 +863,6 @@ var TEXT_MODELS = [
     capabilities: {
       reasoning: true,
       vision: true,
-      tools: true,
-      structuredOutput: true,
       contextWindow: 105e4
     }
   }),
@@ -886,7 +875,7 @@ var TEXT_MODELS = [
       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 }
+    capabilities: { vision: 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", {
@@ -906,8 +895,6 @@ var TEXT_MODELS = [
     capabilities: {
       reasoning: true,
       vision: true,
-      tools: true,
-      structuredOutput: true,
       contextWindow: 1e6
     }
   }),
@@ -922,8 +909,6 @@ var TEXT_MODELS = [
     capabilities: {
       reasoning: true,
       vision: true,
-      tools: true,
-      structuredOutput: true,
       contextWindow: 1e6
     }
   }),
@@ -939,8 +924,6 @@ var TEXT_MODELS = [
     capabilities: {
       reasoning: true,
       vision: true,
-      tools: true,
-      structuredOutput: true,
       contextWindow: 1e6
     }
   }),
@@ -956,8 +939,6 @@ var TEXT_MODELS = [
     capabilities: {
       reasoning: true,
       vision: true,
-      tools: true,
-      structuredOutput: true,
       contextWindow: 2e5
     }
   }),
@@ -971,8 +952,6 @@ var TEXT_MODELS = [
     },
     capabilities: {
       vision: true,
-      tools: true,
-      structuredOutput: true,
       contextWindow: 1e6
     }
   }),
@@ -985,8 +964,6 @@ var TEXT_MODELS = [
     },
     capabilities: {
       vision: true,
-      tools: true,
-      structuredOutput: true,
       contextWindow: 1e6
     }
   }),
@@ -999,8 +976,6 @@ var TEXT_MODELS = [
     },
     capabilities: {
       vision: true,
-      tools: true,
-      structuredOutput: true,
       contextWindow: 1048576
     }
   }),
@@ -1013,8 +988,6 @@ var TEXT_MODELS = [
     },
     capabilities: {
       vision: true,
-      tools: true,
-      structuredOutput: true,
       contextWindow: 1048576
     }
   }),
@@ -1027,8 +1000,6 @@ var TEXT_MODELS = [
     },
     capabilities: {
       vision: true,
-      tools: true,
-      structuredOutput: true,
       contextWindow: 1048576
     }
   }),
@@ -1041,8 +1012,6 @@ var TEXT_MODELS = [
     },
     capabilities: {
       vision: true,
-      tools: true,
-      structuredOutput: true,
       contextWindow: 1048576
     }
   }),
@@ -1055,8 +1024,6 @@ var TEXT_MODELS = [
     },
     capabilities: {
       vision: true,
-      tools: true,
-      structuredOutput: true,
       contextWindow: 1048576
     }
   }),
@@ -1070,8 +1037,6 @@ var TEXT_MODELS = [
     },
     capabilities: {
       vision: true,
-      tools: true,
-      structuredOutput: true,
       contextWindow: 1048576
     }
   }),
@@ -1084,8 +1049,6 @@ var TEXT_MODELS = [
     },
     capabilities: {
       vision: true,
-      tools: true,
-      structuredOutput: true,
       contextWindow: 1048576
     }
   }),
@@ -1104,8 +1067,6 @@ var TEXT_MODELS = [
     },
     capabilities: {
       vision: true,
-      tools: true,
-      structuredOutput: true,
       contextWindow: 1048576
     }
   }),
@@ -1118,8 +1079,6 @@ var TEXT_MODELS = [
     },
     capabilities: {
       vision: true,
-      tools: false,
-      structuredOutput: true,
       contextWindow: 128e3
     }
   }),
@@ -1132,8 +1091,6 @@ var TEXT_MODELS = [
     },
     capabilities: {
       vision: true,
-      tools: false,
-      structuredOutput: true,
       contextWindow: 2e5
     }
   }),
@@ -1146,8 +1103,6 @@ var TEXT_MODELS = [
     },
     capabilities: {
       vision: true,
-      tools: false,
-      structuredOutput: true,
       contextWindow: 128e3
     }
   }),
@@ -1167,8 +1122,6 @@ var TEXT_MODELS = [
     capabilities: {
       reasoning: true,
       vision: true,
-      tools: true,
-      structuredOutput: true,
       contextWindow: 1e6
     }
   }),
@@ -1183,8 +1136,6 @@ var TEXT_MODELS = [
     capabilities: {
       reasoning: true,
       vision: false,
-      tools: true,
-      structuredOutput: true,
       contextWindow: 1e6
     }
   }),
@@ -1198,14 +1149,14 @@ var TEXT_MODELS = [
     },
     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", {
-    exposure: "catalog",
+    lifecycle: "deprecated",
+    replacedBy: "kimi-k3",
+    exposure: "internal",
     display: {
       chat: { temperature: 1 },
       provider: "moonshot",
@@ -1225,8 +1176,6 @@ var TEXT_MODELS = [
     capabilities: {
       reasoning: true,
       vision: true,
-      tools: true,
-      structuredOutput: true,
       contextWindow: 1e6
     }
   }),
@@ -1240,12 +1189,12 @@ var TEXT_MODELS = [
       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 }
+    capabilities: { vision: true, contextWindow: 1e6 }
   })
 ];
 
 // src/catalog/video.ts
-var VIDEO_USAGE = { unit: "seconds", modality: "VIDEO" };
+var VIDEO_USAGE = { unit: "seconds" };
 var SEEDANCE_RATIOS = ["21:9", "16:9", "4:3", "1:1", "3:4", "9:16", "adaptive"];
 var KLING_RATIOS = ["16:9", "9:16", "1:1", "adaptive"];
 var OMNI_RATIOS = ["16:9", "9:16"];
@@ -1419,7 +1368,7 @@ var VIDEO_MODELS = [
     constraints: {
       aspectRatios: [...SEEDANCE_RATIOS],
       resolutions: ["480p", "720p", "1080p", "4k"],
-      durations: [4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15],
+      durations: { min: 4, max: 15 },
       durationPresets: [4, 6, 8, 10, 12, 15],
       references: { images: 9, videos: 3, audios: 3, audioOnly: false },
       ratioIgnoredWith: ["firstFrame"],
@@ -1452,7 +1401,7 @@ var VIDEO_MODELS = [
     constraints: {
       aspectRatios: [...SEEDANCE_RATIOS],
       resolutions: ["480p", "720p"],
-      durations: [4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15],
+      durations: { min: 4, max: 15 },
       references: { images: 9, videos: 3, audios: 3, audioOnly: false },
       ratioIgnoredWith: ["firstFrame"],
       supports: ["generateAudio", "seed", "lastFrame", "watermark"]
@@ -1525,7 +1474,7 @@ 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", slug: "grok" },
+    usage: { unit: "reported_cost_usd", slug: "grok" },
     pricingKind: "video_per_second",
     display: {
       provider: "xai",
@@ -1559,6 +1508,7 @@ var VIDEO_MODELS = [
     pricingKind: "video_per_second",
     display: {
       provider: "kuaishou",
+      logo: "https://cdn.jsdelivr.net/npm/@lobehub/icons-static-png@1.97.0/light/kling-color.png",
       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 }
@@ -1589,6 +1539,7 @@ var VIDEO_MODELS = [
     pricingKind: "video_per_second",
     display: {
       provider: "kuaishou",
+      logo: "https://cdn.jsdelivr.net/npm/@lobehub/icons-static-png@1.97.0/light/kling-color.png",
       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 }
@@ -1622,10 +1573,11 @@ var VIDEO_MODELS = [
       supports: ["generateAudio", "lastFrame"]
     },
     defaults: { durationSeconds: 5, aspectRatio: "16:9", generateAudio: false, mode: "pro" },
-    usage: { unit: "compute_seconds", modality: "VIDEO" },
+    usage: { unit: "compute_seconds" },
     pricingKind: "video_per_second",
     display: {
       provider: "kuaishou",
+      logo: "https://cdn.jsdelivr.net/npm/@lobehub/icons-static-png@1.97.0/light/kling-color.png",
       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: {}
@@ -1655,6 +1607,7 @@ var VIDEO_MODELS = [
     pricingKind: "video_per_second",
     display: {
       provider: "minimax",
+      logo: "https://d31vvfpk2ppdyq.cloudfront.net/assets/image/f4c2e906-28b7-4a4a-b491-df8fef7808ed.png",
       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 }
@@ -1684,6 +1637,7 @@ var VIDEO_MODELS = [
     pricingKind: "video_per_second",
     display: {
       provider: "minimax",
+      logo: "https://d31vvfpk2ppdyq.cloudfront.net/assets/image/f4c2e906-28b7-4a4a-b491-df8fef7808ed.png",
       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 }
@@ -1714,6 +1668,7 @@ var VIDEO_MODELS = [
     pricingKind: "video_per_second",
     display: {
       provider: "minimax",
+      logo: "https://d31vvfpk2ppdyq.cloudfront.net/assets/image/f4c2e906-28b7-4a4a-b491-df8fef7808ed.png",
       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: {}
@@ -1744,6 +1699,7 @@ var VIDEO_MODELS = [
     pricingKind: "video_per_second",
     display: {
       provider: "minimax",
+      logo: "https://d31vvfpk2ppdyq.cloudfront.net/assets/image/f4c2e906-28b7-4a4a-b491-df8fef7808ed.png",
       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: {}
@@ -1769,6 +1725,7 @@ var VIDEO_MODELS = [
     pricingKind: "video_per_second",
     display: {
       provider: "kuaishou",
+      logo: "https://cdn.jsdelivr.net/npm/@lobehub/icons-static-png@1.97.0/light/kling-color.png",
       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 }
@@ -1788,10 +1745,11 @@ var VIDEO_MODELS = [
       supports: ["lastFrame"]
     },
     defaults: { durationSeconds: 6, resolution: "768p" },
-    usage: { unit: "compute_seconds", modality: "VIDEO" },
+    usage: { unit: "compute_seconds" },
     pricingKind: "video_per_second",
     display: {
       provider: "minimax",
+      logo: "https://d31vvfpk2ppdyq.cloudfront.net/assets/image/f4c2e906-28b7-4a4a-b491-df8fef7808ed.png",
       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: {}
@@ -2439,7 +2397,7 @@ function passthroughTextModel(model) {
     exposure: "internal",
     constraints: {},
     defaults: {},
-    usage: { unit: "tokens", modality: "TEXT" },
+    usage: { unit: "tokens" },
     pricingKind: "token"
   };
 }