@shop-os/modelkit 0.1.9 → 0.1.10

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
package/dist/index.cjs CHANGED
@@ -437,6 +437,7 @@ var IMAGE_MODELS = [
437
437
  pricingKind: "image_size",
438
438
  display: {
439
439
  provider: "bytedance",
440
+ logo: "https://d31vvfpk2ppdyq.cloudfront.net/assets/image/3bba5d81-4e57-4e71-911b-cdfa05acca2c_compressed.png",
440
441
  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."
441
442
  },
442
443
  capabilities: { ...IMAGE_GENERATE_EDIT }
@@ -642,6 +643,7 @@ var IMAGE_MODELS = [
642
643
  ];
643
644
 
644
645
  // src/catalog/other.ts
646
+ var GOOGLE_LOGO = "https://d31vvfpk2ppdyq.cloudfront.net/assets/image/192485d2-a76a-4657-9c06-930b2c70badd_compressed.png";
645
647
  var OTHER_MODELS = [
646
648
  {
647
649
  id: "lyria-3-clip-preview",
@@ -662,6 +664,7 @@ var OTHER_MODELS = [
662
664
  pricingKind: "flat",
663
665
  display: {
664
666
  provider: "google",
667
+ logo: GOOGLE_LOGO,
665
668
  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."
666
669
  },
667
670
  capabilities: {}
@@ -684,6 +687,7 @@ var OTHER_MODELS = [
684
687
  pricingKind: "flat",
685
688
  display: {
686
689
  provider: "google",
690
+ logo: GOOGLE_LOGO,
687
691
  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."
688
692
  },
689
693
  capabilities: {}
@@ -909,9 +913,9 @@ var TEXT_MODELS = [
909
913
  }
910
914
  }),
911
915
  text("gpt-5.6-luna", "GPT-5.6 Luna", "gpt-5.6-luna", {
912
- providerOptions: OPENAI_REASONING,
913
916
  exposure: "catalog",
914
917
  display: {
918
+ chat: { providerOptions: OPENAI_REASONING },
915
919
  provider: "openai",
916
920
  costTier: 1,
917
921
  description: "OpenAI GPT-5.6 Luna cost-optimized tier of the GPT-5.6 reasoning family; text and image input, 1.05M-token context."
@@ -925,9 +929,9 @@ var TEXT_MODELS = [
925
929
  }
926
930
  }),
927
931
  text("gpt-5.6-terra", "GPT-5.6 Terra", "gpt-5.6-terra", {
928
- providerOptions: OPENAI_REASONING,
929
932
  exposure: "catalog",
930
933
  display: {
934
+ chat: { providerOptions: OPENAI_REASONING },
931
935
  provider: "openai",
932
936
  costTier: 2,
933
937
  description: "OpenAI GPT-5.6 Terra balanced tier of the GPT-5.6 reasoning family; text and image input, 1.05M-token context."
@@ -941,9 +945,9 @@ var TEXT_MODELS = [
941
945
  }
942
946
  }),
943
947
  text("gpt-5.6-sol", "GPT-5.6 Sol", "gpt-5.6-sol", {
944
- providerOptions: OPENAI_REASONING,
945
948
  exposure: "catalog",
946
949
  display: {
950
+ chat: { providerOptions: OPENAI_REASONING },
947
951
  provider: "openai",
948
952
  costTier: 3,
949
953
  description: "OpenAI GPT-5.6 Sol flagship tier of the GPT-5.6 reasoning family; text and image input, 1.05M-token context."
@@ -969,7 +973,6 @@ var TEXT_MODELS = [
969
973
  }),
970
974
  // Anthropic retired the dated 4.0 ids; callers that still send them get 4.6.
971
975
  text("claude-sonnet-4-6", "Claude Sonnet 4.6", "claude-sonnet-4-6", {
972
- providerOptions: ANTHROPIC_THINKING,
973
976
  exposure: "catalog",
974
977
  aliases: [
975
978
  "claude-sonnet-4.6",
@@ -978,6 +981,7 @@ var TEXT_MODELS = [
978
981
  "claude-sonnet-4-20250514"
979
982
  ],
980
983
  display: {
984
+ chat: { providerOptions: ANTHROPIC_THINKING },
981
985
  provider: "anthropic",
982
986
  costTier: 2,
983
987
  description: "Anthropic Claude Sonnet 4.6 (legacy) balanced model; text and image input, 1M-token context."
@@ -991,9 +995,9 @@ var TEXT_MODELS = [
991
995
  }
992
996
  }),
993
997
  text("claude-sonnet-5", "Claude Sonnet 5", "claude-sonnet-5", {
994
- providerOptions: ANTHROPIC_ADAPTIVE,
995
998
  exposure: "catalog",
996
999
  display: {
1000
+ chat: { providerOptions: ANTHROPIC_ADAPTIVE },
997
1001
  provider: "anthropic",
998
1002
  costTier: 3,
999
1003
  description: "Anthropic Claude Sonnet 5 current balanced model; text and image input, 1M-token context."
@@ -1007,10 +1011,10 @@ var TEXT_MODELS = [
1007
1011
  }
1008
1012
  }),
1009
1013
  text("claude-opus-4-8", "Claude Opus 4.8", "claude-opus-4-8", {
1010
- providerOptions: ANTHROPIC_ADAPTIVE,
1011
1014
  exposure: "catalog",
1012
1015
  aliases: ["claude-opus-4.8", "claude-opus-4", "claude-opus-4.0", "claude-opus-4-20250514"],
1013
1016
  display: {
1017
+ chat: { providerOptions: ANTHROPIC_ADAPTIVE },
1014
1018
  provider: "anthropic",
1015
1019
  costTier: 3,
1016
1020
  description: "Anthropic Claude Opus 4.8 (legacy) frontier model; text and image input, 1M-token context."
@@ -1024,10 +1028,10 @@ var TEXT_MODELS = [
1024
1028
  }
1025
1029
  }),
1026
1030
  text("claude-haiku-4-5", "Claude Haiku 4.5", "claude-haiku-4-5", {
1027
- providerOptions: ANTHROPIC_THINKING,
1028
1031
  exposure: "catalog",
1029
1032
  aliases: ["claude-haiku-4.5", "claude-haiku-4-5-20251001", "claude-haiku-4"],
1030
1033
  display: {
1034
+ chat: { providerOptions: ANTHROPIC_THINKING },
1031
1035
  provider: "anthropic",
1032
1036
  costTier: 1,
1033
1037
  description: "Anthropic Claude Haiku 4.5 fast model with near-frontier intelligence; text and image input, 200K-token context."
@@ -1234,10 +1238,10 @@ var TEXT_MODELS = [
1234
1238
  // Flash rates (https://api-docs.deepseek.com/news/news260910/); the old id
1235
1239
  // stays an alias so stored selections keep resolving.
1236
1240
  text("deepseek-flash", "DeepSeek V4.1 Flash", "deepseek-flash", {
1237
- providerOptions: DEEPSEEK_THINKING,
1238
1241
  aliases: ["deepseek-v4-flash"],
1239
1242
  exposure: "catalog",
1240
1243
  display: {
1244
+ chat: { providerOptions: DEEPSEEK_THINKING },
1241
1245
  provider: "deepseek",
1242
1246
  costTier: 1,
1243
1247
  recommended: true,
@@ -1252,9 +1256,9 @@ var TEXT_MODELS = [
1252
1256
  }
1253
1257
  }),
1254
1258
  text("deepseek-v4-pro", "DeepSeek V4 Pro", "deepseek-v4-pro", {
1255
- providerOptions: DEEPSEEK_THINKING,
1256
1259
  exposure: "catalog",
1257
1260
  display: {
1261
+ chat: { providerOptions: DEEPSEEK_THINKING },
1258
1262
  provider: "deepseek",
1259
1263
  costTier: 2,
1260
1264
  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."
@@ -1284,9 +1288,9 @@ var TEXT_MODELS = [
1284
1288
  }),
1285
1289
  // Moonshot retired kimi-k2.5 on 2026-08-31 and no longer publishes its specs.
1286
1290
  text("kimi-k2.5", "Kimi K2.5", "kimi-k2.5", {
1287
- defaults: { temperature: 1 },
1288
1291
  exposure: "catalog",
1289
1292
  display: {
1293
+ chat: { temperature: 1 },
1290
1294
  provider: "moonshot",
1291
1295
  costTier: 2,
1292
1296
  description: "Moonshot Kimi K2.5; the provider retired the model on 2026-08-31 and calls now return a model-not-found error."
@@ -1294,9 +1298,9 @@ var TEXT_MODELS = [
1294
1298
  capabilities: {}
1295
1299
  }),
1296
1300
  text("kimi-k3", "Kimi K3", "kimi-k3", {
1297
- defaults: { temperature: 1 },
1298
1301
  exposure: "catalog",
1299
1302
  display: {
1303
+ chat: { temperature: 1 },
1300
1304
  provider: "moonshot",
1301
1305
  costTier: 2,
1302
1306
  description: "Moonshot Kimi K3 flagship model with native visual understanding and always-on thinking; text, image, and video input, 1M-token context."
@@ -1333,7 +1337,8 @@ var H3_RATIOS = ["21:9", "16:9", "4:3", "1:1", "3:4", "9:16"];
1333
1337
  var H3_REFERENCE_RATIOS = ["adaptive", ...H3_RATIOS];
1334
1338
  var NEGATIVE_PROMPT = "blur, distort, and low quality";
1335
1339
  var FRAMES_EXCLUDE_REFERENCES = { framesExcludeReferences: true };
1336
- var GOOGLE_LOGO = "https://d31vvfpk2ppdyq.cloudfront.net/assets/image/192485d2-a76a-4657-9c06-930b2c70badd_compressed.png";
1340
+ var BYTEDANCE_LOGO = "https://d31vvfpk2ppdyq.cloudfront.net/assets/image/3bba5d81-4e57-4e71-911b-cdfa05acca2c_compressed.png";
1341
+ var GOOGLE_LOGO2 = "https://d31vvfpk2ppdyq.cloudfront.net/assets/image/192485d2-a76a-4657-9c06-930b2c70badd_compressed.png";
1337
1342
  var VIDEO_MODELS = [
1338
1343
  {
1339
1344
  id: "veo-3.1",
@@ -1362,7 +1367,7 @@ var VIDEO_MODELS = [
1362
1367
  display: {
1363
1368
  provider: "google",
1364
1369
  description: "Google's highest-quality video generation model. Supports up to 4K resolution and 8-second clips with reference image conditioning.",
1365
- logo: GOOGLE_LOGO
1370
+ logo: GOOGLE_LOGO2
1366
1371
  },
1367
1372
  capabilities: { ...FRAMES_EXCLUDE_REFERENCES }
1368
1373
  },
@@ -1454,7 +1459,7 @@ var VIDEO_MODELS = [
1454
1459
  display: {
1455
1460
  provider: "google",
1456
1461
  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.",
1457
- logo: GOOGLE_LOGO
1462
+ logo: GOOGLE_LOGO2
1458
1463
  },
1459
1464
  capabilities: { acceptsSourceVideo: true }
1460
1465
  },
@@ -1479,6 +1484,7 @@ var VIDEO_MODELS = [
1479
1484
  pricingKind: "video_per_second",
1480
1485
  display: {
1481
1486
  provider: "google",
1487
+ logo: GOOGLE_LOGO2,
1482
1488
  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."
1483
1489
  },
1484
1490
  capabilities: { acceptsSourceVideo: true }
@@ -1513,7 +1519,7 @@ var VIDEO_MODELS = [
1513
1519
  display: {
1514
1520
  provider: "bytedance",
1515
1521
  description: "High-quality AI video generation with reference image conditioning. Supports 4-15 second clips with human preprocessing.",
1516
- logo: "https://d31vvfpk2ppdyq.cloudfront.net/assets/image/3bba5d81-4e57-4e71-911b-cdfa05acca2c_compressed.png"
1522
+ logo: BYTEDANCE_LOGO
1517
1523
  },
1518
1524
  capabilities: { acceptsSourceVideo: true, ...FRAMES_EXCLUDE_REFERENCES }
1519
1525
  },
@@ -1544,6 +1550,7 @@ var VIDEO_MODELS = [
1544
1550
  pricingKind: "video_per_second",
1545
1551
  display: {
1546
1552
  provider: "bytedance",
1553
+ logo: BYTEDANCE_LOGO,
1547
1554
  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."
1548
1555
  },
1549
1556
  capabilities: { acceptsSourceVideo: true, ...FRAMES_EXCLUDE_REFERENCES }
@@ -1576,6 +1583,7 @@ var VIDEO_MODELS = [
1576
1583
  pricingKind: "video_per_second",
1577
1584
  display: {
1578
1585
  provider: "bytedance",
1586
+ logo: BYTEDANCE_LOGO,
1579
1587
  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."
1580
1588
  },
1581
1589
  capabilities: { acceptsSourceVideo: true, ...FRAMES_EXCLUDE_REFERENCES }
@@ -1600,7 +1608,7 @@ var VIDEO_MODELS = [
1600
1608
  },
1601
1609
  // No resolution default: xAI picks its own unless the caller asks.
1602
1610
  defaults: { durationSeconds: 8, aspectRatio: "16:9" },
1603
- usage: { unit: "reported_cost_usd", modality: "VIDEO" },
1611
+ usage: { unit: "reported_cost_usd", modality: "VIDEO", slug: "grok" },
1604
1612
  pricingKind: "video_per_second",
1605
1613
  display: {
1606
1614
  provider: "xai",
@@ -1630,7 +1638,7 @@ var VIDEO_MODELS = [
1630
1638
  cfgScale: 0.5,
1631
1639
  negativePrompt: NEGATIVE_PROMPT
1632
1640
  },
1633
- usage: { ...VIDEO_USAGE, slug: "grok" },
1641
+ usage: { ...VIDEO_USAGE },
1634
1642
  pricingKind: "video_per_second",
1635
1643
  display: {
1636
1644
  provider: "kuaishou",