smoltalk 0.4.2 → 0.5.1

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/models.d.ts CHANGED
@@ -1,4 +1,5 @@
1
1
  import { z } from "zod";
2
+ import { type ModelDataBlob, type HostedTool, type HostedToolPrice } from "./modelData.js";
2
3
  export declare const providers: readonly ["ollama", "openai", "openai-responses", "anthropic", "google", "replicate", "modal"];
3
4
  export declare const ProviderSchema: z.ZodEnum<{
4
5
  ollama: "ollama";
@@ -20,6 +21,11 @@ export type BaseModel = {
20
21
  outputTokenCost?: number;
21
22
  disabled?: boolean;
22
23
  costUnit?: "tokens" | "characters" | "minutes";
24
+ knowledge?: string;
25
+ releaseDate?: string;
26
+ lastUpdated?: string;
27
+ family?: string;
28
+ openWeights?: boolean;
23
29
  };
24
30
  export type SpeechToTextModel = BaseModel & {
25
31
  type: "speech-to-text";
@@ -59,6 +65,21 @@ export type TextModel = BaseModel & {
59
65
  /** Whether cryptographic thinking signatures are returned for round-tripping */
60
66
  outputsSignatures?: boolean;
61
67
  };
68
+ modalities?: {
69
+ input: string[];
70
+ output: string[];
71
+ };
72
+ structuredOutput?: boolean;
73
+ temperatureSupported?: boolean;
74
+ inputAudioTokenCost?: number;
75
+ outputAudioTokenCost?: number;
76
+ /** Pricing that applies above a context-size threshold (e.g. Gemini >200k). */
77
+ longContext?: {
78
+ thresholdTokens: number;
79
+ inputTokenCost?: number;
80
+ outputTokenCost?: number;
81
+ cachedInputTokenCost?: number;
82
+ };
62
83
  };
63
84
  export type EmbeddingsModel = {
64
85
  type: "embeddings";
@@ -83,6 +104,17 @@ export declare const textModels: readonly [{
83
104
  readonly cachedInputTokenCost: 0.075;
84
105
  readonly outputTokenCost: 0.6;
85
106
  readonly outputTokensPerSecond: 65;
107
+ readonly modalities: {
108
+ readonly input: readonly ["text", "image", "pdf"];
109
+ readonly output: readonly ["text"];
110
+ };
111
+ readonly knowledge: "2023-09";
112
+ readonly releaseDate: "2024-07-18";
113
+ readonly lastUpdated: "2024-07-18";
114
+ readonly family: "gpt-mini";
115
+ readonly openWeights: false;
116
+ readonly structuredOutput: true;
117
+ readonly temperatureSupported: true;
86
118
  readonly provider: "openai";
87
119
  }, {
88
120
  readonly type: "text";
@@ -94,6 +126,17 @@ export declare const textModels: readonly [{
94
126
  readonly cachedInputTokenCost: 1.25;
95
127
  readonly outputTokenCost: 10;
96
128
  readonly outputTokensPerSecond: 143;
129
+ readonly modalities: {
130
+ readonly input: readonly ["text", "image", "pdf"];
131
+ readonly output: readonly ["text"];
132
+ };
133
+ readonly knowledge: "2023-09";
134
+ readonly releaseDate: "2024-05-13";
135
+ readonly lastUpdated: "2024-08-06";
136
+ readonly family: "gpt";
137
+ readonly openWeights: false;
138
+ readonly structuredOutput: true;
139
+ readonly temperatureSupported: true;
97
140
  readonly provider: "openai";
98
141
  }, {
99
142
  readonly type: "text";
@@ -112,12 +155,23 @@ export declare const textModels: readonly [{
112
155
  readonly outputsThinking: false;
113
156
  readonly outputsSignatures: false;
114
157
  };
158
+ readonly modalities: {
159
+ readonly input: readonly ["text", "image", "pdf"];
160
+ readonly output: readonly ["text"];
161
+ };
162
+ readonly knowledge: "2024-05";
163
+ readonly releaseDate: "2025-04-16";
164
+ readonly lastUpdated: "2025-04-16";
165
+ readonly family: "o";
166
+ readonly openWeights: false;
167
+ readonly structuredOutput: true;
168
+ readonly temperatureSupported: false;
115
169
  readonly provider: "openai";
116
170
  }, {
117
171
  readonly type: "text";
118
172
  readonly modelName: "o3-mini";
119
173
  readonly description: "o3-mini is our most recent small reasoning model, providing high intelligence at the same cost and latency targets of o1-mini. o3-mini also supports key developer features, like Structured Outputs, function calling, Batch API, and more. Like other models in the o-series, it is designed to excel at science, math, and coding tasks. Knowledge cutoff: June 2024.";
120
- readonly maxInputTokens: 500000;
174
+ readonly maxInputTokens: 200000;
121
175
  readonly maxOutputTokens: 100000;
122
176
  readonly inputTokenCost: 1.1;
123
177
  readonly cachedInputTokenCost: 0.55;
@@ -130,6 +184,17 @@ export declare const textModels: readonly [{
130
184
  readonly outputsThinking: false;
131
185
  readonly outputsSignatures: false;
132
186
  };
187
+ readonly modalities: {
188
+ readonly input: readonly ["text"];
189
+ readonly output: readonly ["text"];
190
+ };
191
+ readonly knowledge: "2024-05";
192
+ readonly releaseDate: "2024-12-20";
193
+ readonly lastUpdated: "2025-01-29";
194
+ readonly family: "o-mini";
195
+ readonly openWeights: false;
196
+ readonly structuredOutput: true;
197
+ readonly temperatureSupported: false;
133
198
  readonly provider: "openai";
134
199
  }, {
135
200
  readonly type: "text";
@@ -137,9 +202,9 @@ export declare const textModels: readonly [{
137
202
  readonly description: "Latest small o-series model optimized for fast, effective reasoning with exceptional performance in coding and visual tasks. Knowledge cutoff: June 2024.";
138
203
  readonly maxInputTokens: 200000;
139
204
  readonly maxOutputTokens: 100000;
140
- readonly inputTokenCost: 0.6;
141
- readonly cachedInputTokenCost: 0.3;
142
- readonly outputTokenCost: 2.4;
205
+ readonly inputTokenCost: 1.1;
206
+ readonly cachedInputTokenCost: 0.275;
207
+ readonly outputTokenCost: 4.4;
143
208
  readonly outputTokensPerSecond: 135;
144
209
  readonly reasoning: {
145
210
  readonly levels: readonly ["low", "medium", "high"];
@@ -148,6 +213,17 @@ export declare const textModels: readonly [{
148
213
  readonly outputsThinking: false;
149
214
  readonly outputsSignatures: false;
150
215
  };
216
+ readonly modalities: {
217
+ readonly input: readonly ["text", "image"];
218
+ readonly output: readonly ["text"];
219
+ };
220
+ readonly knowledge: "2024-05";
221
+ readonly releaseDate: "2025-04-16";
222
+ readonly lastUpdated: "2025-04-16";
223
+ readonly family: "o-mini";
224
+ readonly openWeights: false;
225
+ readonly structuredOutput: true;
226
+ readonly temperatureSupported: false;
151
227
  readonly provider: "openai";
152
228
  }, {
153
229
  readonly type: "text";
@@ -162,7 +238,18 @@ export declare const textModels: readonly [{
162
238
  readonly outputsThinking: false;
163
239
  readonly outputsSignatures: false;
164
240
  };
165
- readonly provider: "openai";
241
+ readonly modalities: {
242
+ readonly input: readonly ["text", "image"];
243
+ readonly output: readonly ["text"];
244
+ };
245
+ readonly knowledge: "2024-05";
246
+ readonly releaseDate: "2025-06-10";
247
+ readonly lastUpdated: "2025-06-10";
248
+ readonly family: "o-pro";
249
+ readonly openWeights: false;
250
+ readonly structuredOutput: true;
251
+ readonly temperatureSupported: false;
252
+ readonly provider: "openai-responses";
166
253
  }, {
167
254
  readonly type: "text";
168
255
  readonly modelName: "o1";
@@ -180,6 +267,17 @@ export declare const textModels: readonly [{
180
267
  readonly outputsThinking: false;
181
268
  readonly outputsSignatures: false;
182
269
  };
270
+ readonly modalities: {
271
+ readonly input: readonly ["text", "image", "pdf"];
272
+ readonly output: readonly ["text"];
273
+ };
274
+ readonly knowledge: "2023-09";
275
+ readonly releaseDate: "2024-12-05";
276
+ readonly lastUpdated: "2024-12-05";
277
+ readonly family: "o";
278
+ readonly openWeights: false;
279
+ readonly structuredOutput: true;
280
+ readonly temperatureSupported: false;
183
281
  readonly provider: "openai";
184
282
  }, {
185
283
  readonly type: "text";
@@ -189,6 +287,17 @@ export declare const textModels: readonly [{
189
287
  readonly maxOutputTokens: 4096;
190
288
  readonly inputTokenCost: 10;
191
289
  readonly outputTokenCost: 30;
290
+ readonly modalities: {
291
+ readonly input: readonly ["text", "image"];
292
+ readonly output: readonly ["text"];
293
+ };
294
+ readonly knowledge: "2023-12";
295
+ readonly releaseDate: "2023-11-06";
296
+ readonly lastUpdated: "2024-04-09";
297
+ readonly family: "gpt";
298
+ readonly openWeights: false;
299
+ readonly structuredOutput: false;
300
+ readonly temperatureSupported: true;
192
301
  readonly disabled: true;
193
302
  readonly provider: "openai";
194
303
  }, {
@@ -199,6 +308,17 @@ export declare const textModels: readonly [{
199
308
  readonly maxOutputTokens: 8192;
200
309
  readonly inputTokenCost: 30;
201
310
  readonly outputTokenCost: 60;
311
+ readonly modalities: {
312
+ readonly input: readonly ["text"];
313
+ readonly output: readonly ["text"];
314
+ };
315
+ readonly knowledge: "2023-11";
316
+ readonly releaseDate: "2023-11-06";
317
+ readonly lastUpdated: "2024-04-09";
318
+ readonly family: "gpt";
319
+ readonly openWeights: false;
320
+ readonly structuredOutput: false;
321
+ readonly temperatureSupported: true;
202
322
  readonly disabled: true;
203
323
  readonly provider: "openai";
204
324
  }, {
@@ -208,7 +328,19 @@ export declare const textModels: readonly [{
208
328
  readonly maxInputTokens: 16385;
209
329
  readonly maxOutputTokens: 4096;
210
330
  readonly inputTokenCost: 0.5;
331
+ readonly cachedInputTokenCost: 0;
211
332
  readonly outputTokenCost: 1.5;
333
+ readonly modalities: {
334
+ readonly input: readonly ["text"];
335
+ readonly output: readonly ["text"];
336
+ };
337
+ readonly knowledge: "2021-09-01";
338
+ readonly releaseDate: "2023-03-01";
339
+ readonly lastUpdated: "2023-11-06";
340
+ readonly family: "gpt";
341
+ readonly openWeights: false;
342
+ readonly structuredOutput: false;
343
+ readonly temperatureSupported: true;
212
344
  readonly disabled: true;
213
345
  readonly provider: "openai";
214
346
  }, {
@@ -221,6 +353,17 @@ export declare const textModels: readonly [{
221
353
  readonly cachedInputTokenCost: 0.5;
222
354
  readonly outputTokenCost: 8;
223
355
  readonly outputTokensPerSecond: 105;
356
+ readonly modalities: {
357
+ readonly input: readonly ["text", "image", "pdf"];
358
+ readonly output: readonly ["text"];
359
+ };
360
+ readonly knowledge: "2024-04";
361
+ readonly releaseDate: "2025-04-14";
362
+ readonly lastUpdated: "2025-04-14";
363
+ readonly family: "gpt";
364
+ readonly openWeights: false;
365
+ readonly structuredOutput: true;
366
+ readonly temperatureSupported: true;
224
367
  readonly provider: "openai";
225
368
  }, {
226
369
  readonly type: "text";
@@ -232,6 +375,17 @@ export declare const textModels: readonly [{
232
375
  readonly cachedInputTokenCost: 0.1;
233
376
  readonly outputTokenCost: 1.6;
234
377
  readonly outputTokensPerSecond: 78;
378
+ readonly modalities: {
379
+ readonly input: readonly ["text", "image", "pdf"];
380
+ readonly output: readonly ["text"];
381
+ };
382
+ readonly knowledge: "2024-04";
383
+ readonly releaseDate: "2025-04-14";
384
+ readonly lastUpdated: "2025-04-14";
385
+ readonly family: "gpt-mini";
386
+ readonly openWeights: false;
387
+ readonly structuredOutput: true;
388
+ readonly temperatureSupported: true;
235
389
  readonly provider: "openai";
236
390
  }, {
237
391
  readonly type: "text";
@@ -243,6 +397,17 @@ export declare const textModels: readonly [{
243
397
  readonly cachedInputTokenCost: 0.025;
244
398
  readonly outputTokenCost: 0.4;
245
399
  readonly outputTokensPerSecond: 142;
400
+ readonly modalities: {
401
+ readonly input: readonly ["text", "image"];
402
+ readonly output: readonly ["text"];
403
+ };
404
+ readonly knowledge: "2024-04";
405
+ readonly releaseDate: "2025-04-14";
406
+ readonly lastUpdated: "2025-04-14";
407
+ readonly family: "gpt-nano";
408
+ readonly openWeights: false;
409
+ readonly structuredOutput: true;
410
+ readonly temperatureSupported: true;
246
411
  readonly provider: "openai";
247
412
  }, {
248
413
  readonly type: "text";
@@ -261,7 +426,18 @@ export declare const textModels: readonly [{
261
426
  readonly outputsThinking: false;
262
427
  readonly outputsSignatures: false;
263
428
  };
264
- readonly provider: "openai-responses";
429
+ readonly modalities: {
430
+ readonly input: readonly ["text", "image"];
431
+ readonly output: readonly ["text"];
432
+ };
433
+ readonly knowledge: "2024-09-30";
434
+ readonly releaseDate: "2025-08-07";
435
+ readonly lastUpdated: "2025-08-07";
436
+ readonly family: "gpt";
437
+ readonly openWeights: false;
438
+ readonly structuredOutput: true;
439
+ readonly temperatureSupported: false;
440
+ readonly provider: "openai";
265
441
  }, {
266
442
  readonly type: "text";
267
443
  readonly modelName: "gpt-5-mini";
@@ -279,7 +455,18 @@ export declare const textModels: readonly [{
279
455
  readonly outputsThinking: false;
280
456
  readonly outputsSignatures: false;
281
457
  };
282
- readonly provider: "openai-responses";
458
+ readonly modalities: {
459
+ readonly input: readonly ["text", "image"];
460
+ readonly output: readonly ["text"];
461
+ };
462
+ readonly knowledge: "2024-05-30";
463
+ readonly releaseDate: "2025-08-07";
464
+ readonly lastUpdated: "2025-08-07";
465
+ readonly family: "gpt-mini";
466
+ readonly openWeights: false;
467
+ readonly structuredOutput: true;
468
+ readonly temperatureSupported: false;
469
+ readonly provider: "openai";
283
470
  }, {
284
471
  readonly type: "text";
285
472
  readonly modelName: "gpt-5-nano";
@@ -297,7 +484,18 @@ export declare const textModels: readonly [{
297
484
  readonly outputsThinking: false;
298
485
  readonly outputsSignatures: false;
299
486
  };
300
- readonly provider: "openai-responses";
487
+ readonly modalities: {
488
+ readonly input: readonly ["text", "image"];
489
+ readonly output: readonly ["text"];
490
+ };
491
+ readonly knowledge: "2024-05-30";
492
+ readonly releaseDate: "2025-08-07";
493
+ readonly lastUpdated: "2025-08-07";
494
+ readonly family: "gpt-nano";
495
+ readonly openWeights: false;
496
+ readonly structuredOutput: true;
497
+ readonly temperatureSupported: false;
498
+ readonly provider: "openai";
301
499
  }, {
302
500
  readonly type: "text";
303
501
  readonly modelName: "gpt-5.1";
@@ -314,7 +512,18 @@ export declare const textModels: readonly [{
314
512
  readonly outputsThinking: false;
315
513
  readonly outputsSignatures: false;
316
514
  };
317
- readonly provider: "openai-responses";
515
+ readonly modalities: {
516
+ readonly input: readonly ["text", "image"];
517
+ readonly output: readonly ["text"];
518
+ };
519
+ readonly knowledge: "2024-09-30";
520
+ readonly releaseDate: "2025-11-13";
521
+ readonly lastUpdated: "2025-11-13";
522
+ readonly family: "gpt";
523
+ readonly openWeights: false;
524
+ readonly structuredOutput: true;
525
+ readonly temperatureSupported: false;
526
+ readonly provider: "openai";
318
527
  }, {
319
528
  readonly type: "text";
320
529
  readonly modelName: "gpt-5.2";
@@ -332,7 +541,18 @@ export declare const textModels: readonly [{
332
541
  readonly outputsThinking: false;
333
542
  readonly outputsSignatures: false;
334
543
  };
335
- readonly provider: "openai-responses";
544
+ readonly modalities: {
545
+ readonly input: readonly ["text", "image"];
546
+ readonly output: readonly ["text"];
547
+ };
548
+ readonly knowledge: "2025-08-31";
549
+ readonly releaseDate: "2025-12-11";
550
+ readonly lastUpdated: "2025-12-11";
551
+ readonly family: "gpt";
552
+ readonly openWeights: false;
553
+ readonly structuredOutput: true;
554
+ readonly temperatureSupported: false;
555
+ readonly provider: "openai";
336
556
  }, {
337
557
  readonly type: "text";
338
558
  readonly modelName: "gpt-5.2-pro";
@@ -346,6 +566,17 @@ export declare const textModels: readonly [{
346
566
  readonly outputsThinking: false;
347
567
  readonly outputsSignatures: false;
348
568
  };
569
+ readonly modalities: {
570
+ readonly input: readonly ["text", "image"];
571
+ readonly output: readonly ["text"];
572
+ };
573
+ readonly knowledge: "2025-08-31";
574
+ readonly releaseDate: "2025-12-11";
575
+ readonly lastUpdated: "2025-12-11";
576
+ readonly family: "gpt-pro";
577
+ readonly openWeights: false;
578
+ readonly structuredOutput: false;
579
+ readonly temperatureSupported: false;
349
580
  readonly provider: "openai-responses";
350
581
  }, {
351
582
  readonly type: "text";
@@ -356,6 +587,12 @@ export declare const textModels: readonly [{
356
587
  readonly inputTokenCost: 2.5;
357
588
  readonly cachedInputTokenCost: 0.25;
358
589
  readonly outputTokenCost: 15;
590
+ readonly longContext: {
591
+ readonly inputTokenCost: 5;
592
+ readonly cachedInputTokenCost: 0.5;
593
+ readonly outputTokenCost: 22.5;
594
+ readonly thresholdTokens: 200000;
595
+ };
359
596
  readonly reasoning: {
360
597
  readonly levels: readonly ["none", "low", "medium", "high", "xhigh"];
361
598
  readonly defaultLevel: "none";
@@ -363,7 +600,18 @@ export declare const textModels: readonly [{
363
600
  readonly outputsThinking: false;
364
601
  readonly outputsSignatures: false;
365
602
  };
366
- readonly provider: "openai-responses";
603
+ readonly modalities: {
604
+ readonly input: readonly ["text", "image", "pdf"];
605
+ readonly output: readonly ["text"];
606
+ };
607
+ readonly knowledge: "2025-08-31";
608
+ readonly releaseDate: "2026-03-05";
609
+ readonly lastUpdated: "2026-03-05";
610
+ readonly family: "gpt";
611
+ readonly openWeights: false;
612
+ readonly structuredOutput: true;
613
+ readonly temperatureSupported: false;
614
+ readonly provider: "openai";
367
615
  }, {
368
616
  readonly type: "text";
369
617
  readonly modelName: "gpt-5.4-mini";
@@ -381,7 +629,18 @@ export declare const textModels: readonly [{
381
629
  readonly outputsThinking: false;
382
630
  readonly outputsSignatures: false;
383
631
  };
384
- readonly provider: "openai-responses";
632
+ readonly modalities: {
633
+ readonly input: readonly ["text", "image"];
634
+ readonly output: readonly ["text"];
635
+ };
636
+ readonly knowledge: "2025-08-31";
637
+ readonly releaseDate: "2026-03-17";
638
+ readonly lastUpdated: "2026-03-17";
639
+ readonly family: "gpt-mini";
640
+ readonly openWeights: false;
641
+ readonly structuredOutput: true;
642
+ readonly temperatureSupported: false;
643
+ readonly provider: "openai";
385
644
  }, {
386
645
  readonly type: "text";
387
646
  readonly modelName: "gpt-5.4-nano";
@@ -399,7 +658,18 @@ export declare const textModels: readonly [{
399
658
  readonly outputsThinking: false;
400
659
  readonly outputsSignatures: false;
401
660
  };
402
- readonly provider: "openai-responses";
661
+ readonly modalities: {
662
+ readonly input: readonly ["text", "image"];
663
+ readonly output: readonly ["text"];
664
+ };
665
+ readonly knowledge: "2025-08-31";
666
+ readonly releaseDate: "2026-03-17";
667
+ readonly lastUpdated: "2026-03-17";
668
+ readonly family: "gpt-nano";
669
+ readonly openWeights: false;
670
+ readonly structuredOutput: true;
671
+ readonly temperatureSupported: false;
672
+ readonly provider: "openai";
403
673
  }, {
404
674
  readonly type: "text";
405
675
  readonly modelName: "gpt-5.4-pro";
@@ -408,6 +678,11 @@ export declare const textModels: readonly [{
408
678
  readonly maxOutputTokens: 128000;
409
679
  readonly inputTokenCost: 30;
410
680
  readonly outputTokenCost: 180;
681
+ readonly longContext: {
682
+ readonly inputTokenCost: 60;
683
+ readonly outputTokenCost: 270;
684
+ readonly thresholdTokens: 200000;
685
+ };
411
686
  readonly reasoning: {
412
687
  readonly levels: readonly ["medium", "high", "xhigh"];
413
688
  readonly defaultLevel: "medium";
@@ -415,6 +690,17 @@ export declare const textModels: readonly [{
415
690
  readonly outputsThinking: false;
416
691
  readonly outputsSignatures: false;
417
692
  };
693
+ readonly modalities: {
694
+ readonly input: readonly ["text", "image"];
695
+ readonly output: readonly ["text"];
696
+ };
697
+ readonly knowledge: "2025-08-31";
698
+ readonly releaseDate: "2026-03-05";
699
+ readonly lastUpdated: "2026-03-05";
700
+ readonly family: "gpt-pro";
701
+ readonly openWeights: false;
702
+ readonly structuredOutput: false;
703
+ readonly temperatureSupported: false;
418
704
  readonly provider: "openai-responses";
419
705
  }, {
420
706
  readonly type: "text";
@@ -426,6 +712,12 @@ export declare const textModels: readonly [{
426
712
  readonly cachedInputTokenCost: 0.5;
427
713
  readonly outputTokenCost: 30;
428
714
  readonly outputTokensPerSecond: 66;
715
+ readonly longContext: {
716
+ readonly inputTokenCost: 10;
717
+ readonly cachedInputTokenCost: 1;
718
+ readonly outputTokenCost: 45;
719
+ readonly thresholdTokens: 200000;
720
+ };
429
721
  readonly reasoning: {
430
722
  readonly levels: readonly ["none", "low", "medium", "high", "xhigh"];
431
723
  readonly defaultLevel: "medium";
@@ -433,7 +725,18 @@ export declare const textModels: readonly [{
433
725
  readonly outputsThinking: false;
434
726
  readonly outputsSignatures: false;
435
727
  };
436
- readonly provider: "openai-responses";
728
+ readonly modalities: {
729
+ readonly input: readonly ["text", "image", "pdf"];
730
+ readonly output: readonly ["text"];
731
+ };
732
+ readonly knowledge: "2025-12-01";
733
+ readonly releaseDate: "2026-04-23";
734
+ readonly lastUpdated: "2026-04-23";
735
+ readonly family: "gpt";
736
+ readonly openWeights: false;
737
+ readonly structuredOutput: true;
738
+ readonly temperatureSupported: false;
739
+ readonly provider: "openai";
437
740
  }, {
438
741
  readonly type: "text";
439
742
  readonly modelName: "gpt-5.5-pro";
@@ -442,6 +745,11 @@ export declare const textModels: readonly [{
442
745
  readonly maxOutputTokens: 128000;
443
746
  readonly inputTokenCost: 30;
444
747
  readonly outputTokenCost: 180;
748
+ readonly longContext: {
749
+ readonly inputTokenCost: 60;
750
+ readonly outputTokenCost: 270;
751
+ readonly thresholdTokens: 200000;
752
+ };
445
753
  readonly reasoning: {
446
754
  readonly levels: readonly ["none", "low", "medium", "high", "xhigh"];
447
755
  readonly defaultLevel: "medium";
@@ -449,6 +757,17 @@ export declare const textModels: readonly [{
449
757
  readonly outputsThinking: false;
450
758
  readonly outputsSignatures: false;
451
759
  };
760
+ readonly modalities: {
761
+ readonly input: readonly ["text", "image", "pdf"];
762
+ readonly output: readonly ["text"];
763
+ };
764
+ readonly knowledge: "2025-12-01";
765
+ readonly releaseDate: "2026-04-23";
766
+ readonly lastUpdated: "2026-04-23";
767
+ readonly family: "gpt-pro";
768
+ readonly openWeights: false;
769
+ readonly structuredOutput: true;
770
+ readonly temperatureSupported: false;
452
771
  readonly provider: "openai-responses";
453
772
  }, {
454
773
  readonly type: "text";
@@ -457,8 +776,15 @@ export declare const textModels: readonly [{
457
776
  readonly maxInputTokens: 1048576;
458
777
  readonly maxOutputTokens: 65536;
459
778
  readonly inputTokenCost: 2;
779
+ readonly cachedInputTokenCost: 0.2;
460
780
  readonly outputTokenCost: 12;
461
781
  readonly outputTokensPerSecond: 133;
782
+ readonly longContext: {
783
+ readonly inputTokenCost: 4;
784
+ readonly cachedInputTokenCost: 0.4;
785
+ readonly outputTokenCost: 18;
786
+ readonly thresholdTokens: 200000;
787
+ };
462
788
  readonly reasoning: {
463
789
  readonly levels: readonly ["low", "medium", "high"];
464
790
  readonly defaultLevel: "high";
@@ -466,6 +792,17 @@ export declare const textModels: readonly [{
466
792
  readonly outputsThinking: true;
467
793
  readonly outputsSignatures: true;
468
794
  };
795
+ readonly modalities: {
796
+ readonly input: readonly ["text", "image", "video", "audio", "pdf"];
797
+ readonly output: readonly ["text"];
798
+ };
799
+ readonly knowledge: "2025-01";
800
+ readonly releaseDate: "2026-02-19";
801
+ readonly lastUpdated: "2026-02-19";
802
+ readonly family: "gemini-pro";
803
+ readonly openWeights: false;
804
+ readonly structuredOutput: true;
805
+ readonly temperatureSupported: true;
469
806
  readonly provider: "google";
470
807
  }, {
471
808
  readonly type: "text";
@@ -474,7 +811,25 @@ export declare const textModels: readonly [{
474
811
  readonly maxInputTokens: 1048576;
475
812
  readonly maxOutputTokens: 65536;
476
813
  readonly inputTokenCost: 2;
814
+ readonly cachedInputTokenCost: 0.2;
477
815
  readonly outputTokenCost: 12;
816
+ readonly longContext: {
817
+ readonly inputTokenCost: 4;
818
+ readonly cachedInputTokenCost: 0.4;
819
+ readonly outputTokenCost: 18;
820
+ readonly thresholdTokens: 200000;
821
+ };
822
+ readonly modalities: {
823
+ readonly input: readonly ["text", "image", "video", "audio", "pdf"];
824
+ readonly output: readonly ["text"];
825
+ };
826
+ readonly knowledge: "2025-01";
827
+ readonly releaseDate: "2025-11-18";
828
+ readonly lastUpdated: "2025-11-18";
829
+ readonly family: "gemini-pro";
830
+ readonly openWeights: false;
831
+ readonly structuredOutput: true;
832
+ readonly temperatureSupported: true;
478
833
  readonly disabled: true;
479
834
  readonly provider: "google";
480
835
  }, {
@@ -486,6 +841,7 @@ export declare const textModels: readonly [{
486
841
  readonly inputTokenCost: 1.5;
487
842
  readonly cachedInputTokenCost: 0.15;
488
843
  readonly outputTokenCost: 9;
844
+ readonly inputAudioTokenCost: 1.5;
489
845
  readonly reasoning: {
490
846
  readonly levels: readonly ["minimal", "low", "medium", "high"];
491
847
  readonly defaultLevel: "high";
@@ -493,6 +849,17 @@ export declare const textModels: readonly [{
493
849
  readonly outputsThinking: true;
494
850
  readonly outputsSignatures: true;
495
851
  };
852
+ readonly modalities: {
853
+ readonly input: readonly ["text", "image", "video", "audio", "pdf"];
854
+ readonly output: readonly ["text"];
855
+ };
856
+ readonly knowledge: "2025-01";
857
+ readonly releaseDate: "2026-05-19";
858
+ readonly lastUpdated: "2026-05-19";
859
+ readonly family: "gemini-flash";
860
+ readonly openWeights: false;
861
+ readonly structuredOutput: true;
862
+ readonly temperatureSupported: true;
496
863
  readonly provider: "google";
497
864
  }, {
498
865
  readonly type: "text";
@@ -501,8 +868,10 @@ export declare const textModels: readonly [{
501
868
  readonly maxInputTokens: 1048576;
502
869
  readonly maxOutputTokens: 65536;
503
870
  readonly inputTokenCost: 0.5;
871
+ readonly cachedInputTokenCost: 0.05;
504
872
  readonly outputTokenCost: 3;
505
873
  readonly outputTokensPerSecond: 184;
874
+ readonly inputAudioTokenCost: 1;
506
875
  readonly reasoning: {
507
876
  readonly levels: readonly ["minimal", "low", "medium", "high"];
508
877
  readonly defaultLevel: "high";
@@ -510,6 +879,17 @@ export declare const textModels: readonly [{
510
879
  readonly outputsThinking: true;
511
880
  readonly outputsSignatures: true;
512
881
  };
882
+ readonly modalities: {
883
+ readonly input: readonly ["text", "image", "video", "audio", "pdf"];
884
+ readonly output: readonly ["text"];
885
+ };
886
+ readonly knowledge: "2025-01";
887
+ readonly releaseDate: "2025-12-17";
888
+ readonly lastUpdated: "2025-12-17";
889
+ readonly family: "gemini-flash";
890
+ readonly openWeights: false;
891
+ readonly structuredOutput: true;
892
+ readonly temperatureSupported: true;
513
893
  readonly provider: "google";
514
894
  }, {
515
895
  readonly type: "text";
@@ -518,8 +898,10 @@ export declare const textModels: readonly [{
518
898
  readonly maxInputTokens: 1048576;
519
899
  readonly maxOutputTokens: 65536;
520
900
  readonly inputTokenCost: 0.25;
901
+ readonly cachedInputTokenCost: 0.025;
521
902
  readonly outputTokenCost: 1.5;
522
903
  readonly outputTokensPerSecond: 379;
904
+ readonly inputAudioTokenCost: 0.5;
523
905
  readonly reasoning: {
524
906
  readonly levels: readonly ["minimal", "low", "medium", "high"];
525
907
  readonly defaultLevel: "minimal";
@@ -527,6 +909,17 @@ export declare const textModels: readonly [{
527
909
  readonly outputsThinking: true;
528
910
  readonly outputsSignatures: true;
529
911
  };
912
+ readonly modalities: {
913
+ readonly input: readonly ["text", "image", "video", "audio", "pdf"];
914
+ readonly output: readonly ["text"];
915
+ };
916
+ readonly knowledge: "2025-01";
917
+ readonly releaseDate: "2026-05-07";
918
+ readonly lastUpdated: "2026-05-07";
919
+ readonly family: "gemini-flash-lite";
920
+ readonly openWeights: false;
921
+ readonly structuredOutput: true;
922
+ readonly temperatureSupported: true;
530
923
  readonly provider: "google";
531
924
  }, {
532
925
  readonly type: "text";
@@ -535,24 +928,54 @@ export declare const textModels: readonly [{
535
928
  readonly maxInputTokens: 1048576;
536
929
  readonly maxOutputTokens: 65536;
537
930
  readonly inputTokenCost: 0.25;
931
+ readonly cachedInputTokenCost: 0.025;
538
932
  readonly outputTokenCost: 1.5;
933
+ readonly inputAudioTokenCost: 0.5;
934
+ readonly modalities: {
935
+ readonly input: readonly ["text", "image", "video", "audio", "pdf"];
936
+ readonly output: readonly ["text"];
937
+ };
938
+ readonly knowledge: "2025-01";
939
+ readonly releaseDate: "2026-03-03";
940
+ readonly lastUpdated: "2026-03-03";
941
+ readonly family: "gemini-flash-lite";
942
+ readonly openWeights: false;
943
+ readonly structuredOutput: true;
944
+ readonly temperatureSupported: true;
539
945
  readonly disabled: true;
540
946
  readonly provider: "google";
541
947
  }, {
542
948
  readonly type: "text";
543
949
  readonly modelName: "gemini-2.5-pro";
544
- readonly description: "High-performance Gemini 2.5 model with 2M context window. Adaptive thinking for complex reasoning and coding. Standard pricing for ≤200k tokens ($1.25 input/$10.00 output), higher rates for >200k tokens ($2.50 input/$15.00 output). Batch API: 50% discount.";
545
- readonly maxInputTokens: 2097152;
950
+ readonly description: "High-performance Gemini 2.5 model with 1M context window. Adaptive thinking for complex reasoning and coding. Standard pricing for ≤200k tokens ($1.25 input/$10.00 output), higher rates for >200k tokens ($2.50 input/$15.00 output). Batch API: 50% discount.";
951
+ readonly maxInputTokens: 1048576;
546
952
  readonly maxOutputTokens: 65536;
547
953
  readonly inputTokenCost: 1.25;
548
- readonly cachedInputTokenCost: 0.31;
954
+ readonly cachedInputTokenCost: 0.125;
549
955
  readonly outputTokenCost: 10;
550
956
  readonly outputTokensPerSecond: 145;
957
+ readonly longContext: {
958
+ readonly inputTokenCost: 2.5;
959
+ readonly cachedInputTokenCost: 0.25;
960
+ readonly outputTokenCost: 15;
961
+ readonly thresholdTokens: 200000;
962
+ };
551
963
  readonly reasoning: {
552
964
  readonly canDisable: false;
553
965
  readonly outputsThinking: true;
554
966
  readonly outputsSignatures: true;
555
967
  };
968
+ readonly modalities: {
969
+ readonly input: readonly ["text", "image", "audio", "video", "pdf"];
970
+ readonly output: readonly ["text"];
971
+ };
972
+ readonly knowledge: "2025-01";
973
+ readonly releaseDate: "2025-06-17";
974
+ readonly lastUpdated: "2025-06-17";
975
+ readonly family: "gemini-pro";
976
+ readonly openWeights: false;
977
+ readonly structuredOutput: true;
978
+ readonly temperatureSupported: true;
556
979
  readonly provider: "google";
557
980
  }, {
558
981
  readonly type: "text";
@@ -561,14 +984,26 @@ export declare const textModels: readonly [{
561
984
  readonly maxInputTokens: 1048576;
562
985
  readonly maxOutputTokens: 65536;
563
986
  readonly inputTokenCost: 0.3;
564
- readonly cachedInputTokenCost: 0.075;
987
+ readonly cachedInputTokenCost: 0.03;
565
988
  readonly outputTokenCost: 2.5;
566
989
  readonly outputTokensPerSecond: 245;
990
+ readonly inputAudioTokenCost: 1;
567
991
  readonly reasoning: {
568
992
  readonly canDisable: true;
569
993
  readonly outputsThinking: true;
570
994
  readonly outputsSignatures: true;
571
995
  };
996
+ readonly modalities: {
997
+ readonly input: readonly ["text", "image", "audio", "video", "pdf"];
998
+ readonly output: readonly ["text"];
999
+ };
1000
+ readonly knowledge: "2025-01";
1001
+ readonly releaseDate: "2025-06-17";
1002
+ readonly lastUpdated: "2025-06-17";
1003
+ readonly family: "gemini-flash";
1004
+ readonly openWeights: false;
1005
+ readonly structuredOutput: true;
1006
+ readonly temperatureSupported: true;
572
1007
  readonly provider: "google";
573
1008
  }, {
574
1009
  readonly type: "text";
@@ -577,14 +1012,26 @@ export declare const textModels: readonly [{
577
1012
  readonly maxInputTokens: 1048576;
578
1013
  readonly maxOutputTokens: 65536;
579
1014
  readonly inputTokenCost: 0.1;
580
- readonly cachedInputTokenCost: 0.025;
1015
+ readonly cachedInputTokenCost: 0.01;
581
1016
  readonly outputTokenCost: 0.4;
582
1017
  readonly outputTokensPerSecond: 400;
1018
+ readonly inputAudioTokenCost: 0.3;
583
1019
  readonly reasoning: {
584
1020
  readonly canDisable: true;
585
1021
  readonly outputsThinking: true;
586
1022
  readonly outputsSignatures: false;
587
1023
  };
1024
+ readonly modalities: {
1025
+ readonly input: readonly ["text", "image", "audio", "video", "pdf"];
1026
+ readonly output: readonly ["text"];
1027
+ };
1028
+ readonly knowledge: "2025-01";
1029
+ readonly releaseDate: "2025-06-17";
1030
+ readonly lastUpdated: "2025-06-17";
1031
+ readonly family: "gemini-flash-lite";
1032
+ readonly openWeights: false;
1033
+ readonly structuredOutput: true;
1034
+ readonly temperatureSupported: true;
588
1035
  readonly provider: "google";
589
1036
  }, {
590
1037
  readonly type: "text";
@@ -593,8 +1040,20 @@ export declare const textModels: readonly [{
593
1040
  readonly maxInputTokens: 1048576;
594
1041
  readonly maxOutputTokens: 8192;
595
1042
  readonly inputTokenCost: 0.1;
1043
+ readonly cachedInputTokenCost: 0.025;
596
1044
  readonly outputTokenCost: 0.4;
597
1045
  readonly outputTokensPerSecond: 213;
1046
+ readonly modalities: {
1047
+ readonly input: readonly ["text", "image", "audio", "video", "pdf"];
1048
+ readonly output: readonly ["text"];
1049
+ };
1050
+ readonly knowledge: "2024-06";
1051
+ readonly releaseDate: "2024-12-11";
1052
+ readonly lastUpdated: "2024-12-11";
1053
+ readonly family: "gemini-flash";
1054
+ readonly openWeights: false;
1055
+ readonly structuredOutput: true;
1056
+ readonly temperatureSupported: true;
598
1057
  readonly disabled: true;
599
1058
  readonly provider: "google";
600
1059
  }, {
@@ -615,6 +1074,17 @@ export declare const textModels: readonly [{
615
1074
  readonly maxOutputTokens: 8192;
616
1075
  readonly inputTokenCost: 0.075;
617
1076
  readonly outputTokenCost: 0.3;
1077
+ readonly modalities: {
1078
+ readonly input: readonly ["text", "image", "audio", "video", "pdf"];
1079
+ readonly output: readonly ["text"];
1080
+ };
1081
+ readonly knowledge: "2024-06";
1082
+ readonly releaseDate: "2024-12-11";
1083
+ readonly lastUpdated: "2024-12-11";
1084
+ readonly family: "gemini-flash-lite";
1085
+ readonly openWeights: false;
1086
+ readonly structuredOutput: true;
1087
+ readonly temperatureSupported: true;
618
1088
  readonly disabled: true;
619
1089
  readonly provider: "google";
620
1090
  }, {
@@ -660,6 +1130,7 @@ export declare const textModels: readonly [{
660
1130
  readonly maxOutputTokens: 128000;
661
1131
  readonly inputTokenCost: 5;
662
1132
  readonly cachedInputTokenCost: 0.5;
1133
+ readonly cacheCreationInputTokenCost: 6.25;
663
1134
  readonly outputTokenCost: 25;
664
1135
  readonly reasoning: {
665
1136
  readonly thinkingStyle: "adaptive";
@@ -667,6 +1138,15 @@ export declare const textModels: readonly [{
667
1138
  readonly outputsThinking: true;
668
1139
  readonly outputsSignatures: true;
669
1140
  };
1141
+ readonly modalities: {
1142
+ readonly input: readonly ["text", "image", "pdf"];
1143
+ readonly output: readonly ["text"];
1144
+ };
1145
+ readonly releaseDate: "2026-05-28";
1146
+ readonly lastUpdated: "2026-05-28";
1147
+ readonly family: "claude-opus";
1148
+ readonly openWeights: false;
1149
+ readonly temperatureSupported: false;
670
1150
  readonly provider: "anthropic";
671
1151
  }, {
672
1152
  readonly type: "text";
@@ -685,6 +1165,16 @@ export declare const textModels: readonly [{
685
1165
  readonly outputsThinking: true;
686
1166
  readonly outputsSignatures: true;
687
1167
  };
1168
+ readonly modalities: {
1169
+ readonly input: readonly ["text", "image", "pdf"];
1170
+ readonly output: readonly ["text"];
1171
+ };
1172
+ readonly knowledge: "2026-01-31";
1173
+ readonly releaseDate: "2026-04-16";
1174
+ readonly lastUpdated: "2026-04-16";
1175
+ readonly family: "claude-opus";
1176
+ readonly openWeights: false;
1177
+ readonly temperatureSupported: false;
688
1178
  readonly provider: "anthropic";
689
1179
  }, {
690
1180
  readonly type: "text";
@@ -703,6 +1193,16 @@ export declare const textModels: readonly [{
703
1193
  readonly outputsThinking: true;
704
1194
  readonly outputsSignatures: true;
705
1195
  };
1196
+ readonly modalities: {
1197
+ readonly input: readonly ["text", "image", "pdf"];
1198
+ readonly output: readonly ["text"];
1199
+ };
1200
+ readonly knowledge: "2025-05-31";
1201
+ readonly releaseDate: "2026-02-05";
1202
+ readonly lastUpdated: "2026-03-13";
1203
+ readonly family: "claude-opus";
1204
+ readonly openWeights: false;
1205
+ readonly temperatureSupported: true;
706
1206
  readonly provider: "anthropic";
707
1207
  }, {
708
1208
  readonly type: "text";
@@ -721,6 +1221,16 @@ export declare const textModels: readonly [{
721
1221
  readonly outputsThinking: true;
722
1222
  readonly outputsSignatures: true;
723
1223
  };
1224
+ readonly modalities: {
1225
+ readonly input: readonly ["text", "image", "pdf"];
1226
+ readonly output: readonly ["text"];
1227
+ };
1228
+ readonly knowledge: "2025-08-31";
1229
+ readonly releaseDate: "2026-02-17";
1230
+ readonly lastUpdated: "2026-03-13";
1231
+ readonly family: "claude-sonnet";
1232
+ readonly openWeights: false;
1233
+ readonly temperatureSupported: true;
724
1234
  readonly provider: "anthropic";
725
1235
  }, {
726
1236
  readonly type: "text";
@@ -739,6 +1249,16 @@ export declare const textModels: readonly [{
739
1249
  readonly outputsThinking: true;
740
1250
  readonly outputsSignatures: true;
741
1251
  };
1252
+ readonly modalities: {
1253
+ readonly input: readonly ["text", "image", "pdf"];
1254
+ readonly output: readonly ["text"];
1255
+ };
1256
+ readonly knowledge: "2025-02-28";
1257
+ readonly releaseDate: "2025-10-15";
1258
+ readonly lastUpdated: "2025-10-15";
1259
+ readonly family: "claude-haiku";
1260
+ readonly openWeights: false;
1261
+ readonly temperatureSupported: true;
742
1262
  readonly provider: "anthropic";
743
1263
  }, {
744
1264
  readonly type: "text";
@@ -768,6 +1288,112 @@ export declare const textModels: readonly [{
768
1288
  readonly outputTokensPerSecond: 66;
769
1289
  readonly disabled: true;
770
1290
  readonly provider: "anthropic";
1291
+ }, {
1292
+ readonly type: "text";
1293
+ readonly modelName: "claude-haiku-4-5";
1294
+ readonly description: "Undated alias for claude-haiku-4-5-20251001. The fastest Claude model with near-frontier intelligence.";
1295
+ readonly maxInputTokens: 200000;
1296
+ readonly maxOutputTokens: 64000;
1297
+ readonly inputTokenCost: 1;
1298
+ readonly cachedInputTokenCost: 0.1;
1299
+ readonly cacheCreationInputTokenCost: 1.25;
1300
+ readonly outputTokenCost: 5;
1301
+ readonly reasoning: {
1302
+ readonly thinkingStyle: "budget";
1303
+ readonly canDisable: true;
1304
+ readonly outputsThinking: true;
1305
+ readonly outputsSignatures: true;
1306
+ };
1307
+ readonly modalities: {
1308
+ readonly input: readonly ["text", "image", "pdf"];
1309
+ readonly output: readonly ["text"];
1310
+ };
1311
+ readonly knowledge: "2025-02-28";
1312
+ readonly releaseDate: "2025-10-15";
1313
+ readonly lastUpdated: "2025-10-15";
1314
+ readonly family: "claude-haiku";
1315
+ readonly openWeights: false;
1316
+ readonly temperatureSupported: true;
1317
+ readonly provider: "anthropic";
1318
+ }, {
1319
+ readonly type: "text";
1320
+ readonly modelName: "claude-opus-4-5";
1321
+ readonly description: "Claude Opus 4.5 — earlier Opus generation. Prefer claude-opus-4-8.";
1322
+ readonly maxInputTokens: 200000;
1323
+ readonly maxOutputTokens: 64000;
1324
+ readonly inputTokenCost: 5;
1325
+ readonly cachedInputTokenCost: 0.5;
1326
+ readonly cacheCreationInputTokenCost: 6.25;
1327
+ readonly outputTokenCost: 25;
1328
+ readonly reasoning: {
1329
+ readonly thinkingStyle: "budget";
1330
+ readonly canDisable: true;
1331
+ readonly outputsThinking: true;
1332
+ readonly outputsSignatures: true;
1333
+ };
1334
+ readonly modalities: {
1335
+ readonly input: readonly ["text", "image", "pdf"];
1336
+ readonly output: readonly ["text"];
1337
+ };
1338
+ readonly knowledge: "2025-03-31";
1339
+ readonly releaseDate: "2025-11-24";
1340
+ readonly lastUpdated: "2025-11-24";
1341
+ readonly family: "claude-opus";
1342
+ readonly openWeights: false;
1343
+ readonly temperatureSupported: true;
1344
+ readonly provider: "anthropic";
1345
+ }, {
1346
+ readonly type: "text";
1347
+ readonly modelName: "claude-sonnet-4-5";
1348
+ readonly description: "Claude Sonnet 4.5 — earlier Sonnet generation. Prefer claude-sonnet-4-6.";
1349
+ readonly maxInputTokens: 200000;
1350
+ readonly maxOutputTokens: 64000;
1351
+ readonly inputTokenCost: 3;
1352
+ readonly cachedInputTokenCost: 0.3;
1353
+ readonly cacheCreationInputTokenCost: 3.75;
1354
+ readonly outputTokenCost: 15;
1355
+ readonly reasoning: {
1356
+ readonly thinkingStyle: "budget";
1357
+ readonly canDisable: true;
1358
+ readonly outputsThinking: true;
1359
+ readonly outputsSignatures: true;
1360
+ };
1361
+ readonly modalities: {
1362
+ readonly input: readonly ["text", "image", "pdf"];
1363
+ readonly output: readonly ["text"];
1364
+ };
1365
+ readonly knowledge: "2025-07-31";
1366
+ readonly releaseDate: "2025-09-29";
1367
+ readonly lastUpdated: "2025-09-29";
1368
+ readonly family: "claude-sonnet";
1369
+ readonly openWeights: false;
1370
+ readonly temperatureSupported: true;
1371
+ readonly provider: "anthropic";
1372
+ }, {
1373
+ readonly type: "text";
1374
+ readonly modelName: "gpt-5-pro";
1375
+ readonly description: "GPT-5 Pro uses more compute for complex reasoning tasks. Available via the Responses API.";
1376
+ readonly maxInputTokens: 400000;
1377
+ readonly maxOutputTokens: 272000;
1378
+ readonly inputTokenCost: 15;
1379
+ readonly outputTokenCost: 120;
1380
+ readonly reasoning: {
1381
+ readonly canDisable: false;
1382
+ readonly outputsThinking: false;
1383
+ readonly outputsSignatures: false;
1384
+ };
1385
+ readonly modalities: {
1386
+ readonly input: readonly ["text", "image"];
1387
+ readonly output: readonly ["text"];
1388
+ };
1389
+ readonly knowledge: "2024-09-30";
1390
+ readonly releaseDate: "2025-10-06";
1391
+ readonly lastUpdated: "2025-10-06";
1392
+ readonly family: "gpt-pro";
1393
+ readonly openWeights: false;
1394
+ readonly structuredOutput: true;
1395
+ readonly temperatureSupported: false;
1396
+ readonly provider: "openai-responses";
771
1397
  }];
772
1398
  export declare const imageModels: readonly [{
773
1399
  readonly type: "image";
@@ -816,751 +1442,23 @@ export type ImageModelName = (typeof imageModels)[number]["modelName"];
816
1442
  export type SpeechToTextModelName = (typeof speechToTextModels)[number]["modelName"];
817
1443
  export type EmbeddingsModelName = (typeof embeddingsModels)[number]["modelName"];
818
1444
  export type ModelName = string;
1445
+ export declare const hostedTools: HostedTool[];
819
1446
  export declare const registeredTextModels: TextModel[];
820
1447
  export declare function registerTextModel(model: Omit<TextModel, "type"> & {
821
1448
  type?: "text";
822
1449
  }): void;
823
- export declare function getModel(modelName: ModelName): TextModel | EmbeddingsModel | {
824
- readonly type: "speech-to-text";
825
- readonly modelName: "whisper-web";
826
- readonly perMinuteCost: 0.006;
827
- readonly provider: "openai";
828
- } | {
829
- readonly type: "text";
830
- readonly modelName: "gpt-4o-mini";
831
- readonly description: "GPT-4o mini ('o' for 'omni') is a fast, affordable small model for focused tasks. It accepts both text and image inputs, and produces text outputs (including Structured Outputs). It is ideal for fine-tuning, and model outputs from a larger model like GPT-4o can be distilled to GPT-4o-mini to produce similar results at lower cost and latency. Knowledge cutoff: July 2025.";
832
- readonly maxInputTokens: 128000;
833
- readonly maxOutputTokens: 16384;
834
- readonly inputTokenCost: 0.15;
835
- readonly cachedInputTokenCost: 0.075;
836
- readonly outputTokenCost: 0.6;
837
- readonly outputTokensPerSecond: 65;
838
- readonly provider: "openai";
839
- } | {
840
- readonly type: "text";
841
- readonly modelName: "gpt-4o";
842
- readonly description: "GPT-4o ('o' for 'omni') is our versatile, high-intelligence flagship model. It accepts both text and image inputs, and produces text outputs (including Structured Outputs). Knowledge cutoff: April 2024.";
843
- readonly maxInputTokens: 128000;
844
- readonly maxOutputTokens: 16384;
845
- readonly inputTokenCost: 2.5;
846
- readonly cachedInputTokenCost: 1.25;
847
- readonly outputTokenCost: 10;
848
- readonly outputTokensPerSecond: 143;
849
- readonly provider: "openai";
850
- } | {
851
- readonly type: "text";
852
- readonly modelName: "o3";
853
- readonly description: "o3 is a reasoning model that sets a new standard for math, science, coding, visual reasoning tasks, and technical writing. Part of the o-series of reasoning models. Knowledge cutoff: June 2024.";
854
- readonly maxInputTokens: 200000;
855
- readonly maxOutputTokens: 100000;
856
- readonly inputTokenCost: 2;
857
- readonly cachedInputTokenCost: 0.5;
858
- readonly outputTokenCost: 8;
859
- readonly outputTokensPerSecond: 94;
860
- readonly reasoning: {
861
- readonly levels: readonly ["low", "medium", "high"];
862
- readonly defaultLevel: "medium";
863
- readonly canDisable: false;
864
- readonly outputsThinking: false;
865
- readonly outputsSignatures: false;
866
- };
867
- readonly provider: "openai";
868
- } | {
869
- readonly type: "text";
870
- readonly modelName: "o3-mini";
871
- readonly description: "o3-mini is our most recent small reasoning model, providing high intelligence at the same cost and latency targets of o1-mini. o3-mini also supports key developer features, like Structured Outputs, function calling, Batch API, and more. Like other models in the o-series, it is designed to excel at science, math, and coding tasks. Knowledge cutoff: June 2024.";
872
- readonly maxInputTokens: 500000;
873
- readonly maxOutputTokens: 100000;
874
- readonly inputTokenCost: 1.1;
875
- readonly cachedInputTokenCost: 0.55;
876
- readonly outputTokenCost: 4.4;
877
- readonly outputTokensPerSecond: 214;
878
- readonly reasoning: {
879
- readonly levels: readonly ["low", "medium", "high"];
880
- readonly defaultLevel: "medium";
881
- readonly canDisable: false;
882
- readonly outputsThinking: false;
883
- readonly outputsSignatures: false;
884
- };
885
- readonly provider: "openai";
886
- } | {
887
- readonly type: "text";
888
- readonly modelName: "o4-mini";
889
- readonly description: "Latest small o-series model optimized for fast, effective reasoning with exceptional performance in coding and visual tasks. Knowledge cutoff: June 2024.";
890
- readonly maxInputTokens: 200000;
891
- readonly maxOutputTokens: 100000;
892
- readonly inputTokenCost: 0.6;
893
- readonly cachedInputTokenCost: 0.3;
894
- readonly outputTokenCost: 2.4;
895
- readonly outputTokensPerSecond: 135;
896
- readonly reasoning: {
897
- readonly levels: readonly ["low", "medium", "high"];
898
- readonly defaultLevel: "medium";
899
- readonly canDisable: false;
900
- readonly outputsThinking: false;
901
- readonly outputsSignatures: false;
902
- };
903
- readonly provider: "openai";
904
- } | {
905
- readonly type: "text";
906
- readonly modelName: "o3-pro";
907
- readonly description: "o3-pro uses more compute for complex reasoning tasks. Available via Responses API only. Requests may take several minutes. Knowledge cutoff: June 2024.";
908
- readonly maxInputTokens: 200000;
909
- readonly maxOutputTokens: 100000;
910
- readonly inputTokenCost: 20;
911
- readonly outputTokenCost: 80;
912
- readonly reasoning: {
913
- readonly canDisable: false;
914
- readonly outputsThinking: false;
915
- readonly outputsSignatures: false;
916
- };
917
- readonly provider: "openai";
918
- } | {
919
- readonly type: "text";
920
- readonly modelName: "o1";
921
- readonly description: "o1 is a reasoning model designed to excel at complex reasoning tasks including science, math, and coding. The knowledge cutoff for o1 models is October, 2023.";
922
- readonly maxInputTokens: 200000;
923
- readonly maxOutputTokens: 100000;
924
- readonly inputTokenCost: 15;
925
- readonly cachedInputTokenCost: 7.5;
926
- readonly outputTokenCost: 60;
927
- readonly outputTokensPerSecond: 100;
928
- readonly reasoning: {
929
- readonly levels: readonly ["low", "medium", "high"];
930
- readonly defaultLevel: "medium";
931
- readonly canDisable: false;
932
- readonly outputsThinking: false;
933
- readonly outputsSignatures: false;
934
- };
935
- readonly provider: "openai";
936
- } | {
937
- readonly type: "text";
938
- readonly modelName: "gpt-4-turbo";
939
- readonly description: "GPT-4 is an older version of a high-intelligence GPT model, usable in Chat Completions. Learn more in the text generation guide. The knowledge cutoff for the latest GPT-4 Turbo version is December, 2023.";
940
- readonly maxInputTokens: 128000;
941
- readonly maxOutputTokens: 4096;
942
- readonly inputTokenCost: 10;
943
- readonly outputTokenCost: 30;
944
- readonly disabled: true;
945
- readonly provider: "openai";
946
- } | {
947
- readonly type: "text";
948
- readonly modelName: "gpt-4";
949
- readonly description: "GPT-4 is an older version of a high-intelligence GPT model, usable in Chat Completions. Learn more in the text generation guide. The knowledge cutoff for the latest GPT-4 Turbo version is December, 2023.";
950
- readonly maxInputTokens: 8192;
951
- readonly maxOutputTokens: 8192;
952
- readonly inputTokenCost: 30;
953
- readonly outputTokenCost: 60;
954
- readonly disabled: true;
955
- readonly provider: "openai";
956
- } | {
957
- readonly type: "text";
958
- readonly modelName: "gpt-3.5-turbo";
959
- readonly description: "GPT-3.5 Turbo models can understand and generate natural language or code and have been optimized for chat using the Chat Completions API but work well for non-chat tasks as well. gpt-4o-mini should be used in place of gpt-3.5-turbo, as it is cheaper, more capable, multimodal, and just as fast.";
960
- readonly maxInputTokens: 16385;
961
- readonly maxOutputTokens: 4096;
962
- readonly inputTokenCost: 0.5;
963
- readonly outputTokenCost: 1.5;
964
- readonly disabled: true;
965
- readonly provider: "openai";
966
- } | {
967
- readonly type: "text";
968
- readonly modelName: "gpt-4.1";
969
- readonly description: "GPT-4.1 excels at instruction following and tool calling with 1M token context window. Knowledge cutoff: June 2024.";
970
- readonly maxInputTokens: 1047576;
971
- readonly maxOutputTokens: 32768;
972
- readonly inputTokenCost: 2;
973
- readonly cachedInputTokenCost: 0.5;
974
- readonly outputTokenCost: 8;
975
- readonly outputTokensPerSecond: 105;
976
- readonly provider: "openai";
977
- } | {
978
- readonly type: "text";
979
- readonly modelName: "gpt-4.1-mini";
980
- readonly description: "GPT-4.1 mini excels at instruction following and tool calling with 1M token context window and low latency. Knowledge cutoff: June 2024.";
981
- readonly maxInputTokens: 1047576;
982
- readonly maxOutputTokens: 32768;
983
- readonly inputTokenCost: 0.4;
984
- readonly cachedInputTokenCost: 0.1;
985
- readonly outputTokenCost: 1.6;
986
- readonly outputTokensPerSecond: 78;
987
- readonly provider: "openai";
988
- } | {
989
- readonly type: "text";
990
- readonly modelName: "gpt-4.1-nano";
991
- readonly description: "GPT-4.1 nano is the fastest and most affordable GPT-4.1 variant with 1M token context window. Knowledge cutoff: June 2024.";
992
- readonly maxInputTokens: 1047576;
993
- readonly maxOutputTokens: 32768;
994
- readonly inputTokenCost: 0.1;
995
- readonly cachedInputTokenCost: 0.025;
996
- readonly outputTokenCost: 0.4;
997
- readonly outputTokensPerSecond: 142;
998
- readonly provider: "openai";
999
- } | {
1000
- readonly type: "text";
1001
- readonly modelName: "gpt-5";
1002
- readonly description: "GPT-5 is a frontier reasoning model with 400K context window. Supports reasoning tokens. Knowledge cutoff: September 2024.";
1003
- readonly maxInputTokens: 400000;
1004
- readonly maxOutputTokens: 128000;
1005
- readonly inputTokenCost: 1.25;
1006
- readonly cachedInputTokenCost: 0.125;
1007
- readonly outputTokenCost: 10;
1008
- readonly outputTokensPerSecond: 72;
1009
- readonly reasoning: {
1010
- readonly levels: readonly ["minimal", "low", "medium", "high"];
1011
- readonly defaultLevel: "medium";
1012
- readonly canDisable: false;
1013
- readonly outputsThinking: false;
1014
- readonly outputsSignatures: false;
1015
- };
1016
- readonly provider: "openai-responses";
1017
- } | {
1018
- readonly type: "text";
1019
- readonly modelName: "gpt-5-mini";
1020
- readonly description: "GPT-5 mini is a faster, more cost-efficient version of GPT-5 with 400K context window. Knowledge cutoff: May 2024.";
1021
- readonly maxInputTokens: 400000;
1022
- readonly maxOutputTokens: 128000;
1023
- readonly inputTokenCost: 0.25;
1024
- readonly cachedInputTokenCost: 0.025;
1025
- readonly outputTokenCost: 2;
1026
- readonly outputTokensPerSecond: 69;
1027
- readonly reasoning: {
1028
- readonly levels: readonly ["minimal", "low", "medium", "high"];
1029
- readonly defaultLevel: "medium";
1030
- readonly canDisable: false;
1031
- readonly outputsThinking: false;
1032
- readonly outputsSignatures: false;
1033
- };
1034
- readonly provider: "openai-responses";
1035
- } | {
1036
- readonly type: "text";
1037
- readonly modelName: "gpt-5-nano";
1038
- readonly description: "GPT-5 nano is the fastest and most affordable GPT-5 variant with 400K context window. Knowledge cutoff: May 2024.";
1039
- readonly maxInputTokens: 400000;
1040
- readonly maxOutputTokens: 128000;
1041
- readonly inputTokenCost: 0.05;
1042
- readonly cachedInputTokenCost: 0.005;
1043
- readonly outputTokenCost: 0.4;
1044
- readonly outputTokensPerSecond: 140;
1045
- readonly reasoning: {
1046
- readonly levels: readonly ["minimal", "low", "medium", "high"];
1047
- readonly defaultLevel: "medium";
1048
- readonly canDisable: false;
1049
- readonly outputsThinking: false;
1050
- readonly outputsSignatures: false;
1051
- };
1052
- readonly provider: "openai-responses";
1053
- } | {
1054
- readonly type: "text";
1055
- readonly modelName: "gpt-5.1";
1056
- readonly description: "GPT-5.1 is the flagship model for coding and agentic tasks with configurable reasoning effort. 400K context window. Knowledge cutoff: September 2024.";
1057
- readonly maxInputTokens: 400000;
1058
- readonly maxOutputTokens: 128000;
1059
- readonly inputTokenCost: 1.25;
1060
- readonly cachedInputTokenCost: 0.125;
1061
- readonly outputTokenCost: 10;
1062
- readonly reasoning: {
1063
- readonly levels: readonly ["none", "low", "medium", "high"];
1064
- readonly defaultLevel: "none";
1065
- readonly canDisable: true;
1066
- readonly outputsThinking: false;
1067
- readonly outputsSignatures: false;
1068
- };
1069
- readonly provider: "openai-responses";
1070
- } | {
1071
- readonly type: "text";
1072
- readonly modelName: "gpt-5.2";
1073
- readonly description: "GPT-5.2 is the flagship model for coding and agentic tasks across industries. 400K context window. Knowledge cutoff: August 2025.";
1074
- readonly maxInputTokens: 400000;
1075
- readonly maxOutputTokens: 128000;
1076
- readonly inputTokenCost: 1.75;
1077
- readonly cachedInputTokenCost: 0.175;
1078
- readonly outputTokenCost: 14;
1079
- readonly outputTokensPerSecond: 61;
1080
- readonly reasoning: {
1081
- readonly levels: readonly ["none", "low", "medium", "high"];
1082
- readonly defaultLevel: "none";
1083
- readonly canDisable: true;
1084
- readonly outputsThinking: false;
1085
- readonly outputsSignatures: false;
1086
- };
1087
- readonly provider: "openai-responses";
1088
- } | {
1089
- readonly type: "text";
1090
- readonly modelName: "gpt-5.2-pro";
1091
- readonly description: "GPT-5.2 Pro uses more compute for complex reasoning tasks. 400K context window. Knowledge cutoff: August 2025.";
1092
- readonly maxInputTokens: 400000;
1093
- readonly maxOutputTokens: 128000;
1094
- readonly inputTokenCost: 21;
1095
- readonly outputTokenCost: 168;
1096
- readonly reasoning: {
1097
- readonly canDisable: false;
1098
- readonly outputsThinking: false;
1099
- readonly outputsSignatures: false;
1100
- };
1101
- readonly provider: "openai-responses";
1102
- } | {
1103
- readonly type: "text";
1104
- readonly modelName: "gpt-5.4";
1105
- readonly description: "GPT-5.4 is the most capable and efficient frontier model for complex professional work. 1M context window, state-of-the-art coding and tool use. Standard pricing for ≤272K tokens, 2x input/1.5x output for >272K. Knowledge cutoff: August 2025.";
1106
- readonly maxInputTokens: 1050000;
1107
- readonly maxOutputTokens: 128000;
1108
- readonly inputTokenCost: 2.5;
1109
- readonly cachedInputTokenCost: 0.25;
1110
- readonly outputTokenCost: 15;
1111
- readonly reasoning: {
1112
- readonly levels: readonly ["none", "low", "medium", "high", "xhigh"];
1113
- readonly defaultLevel: "none";
1114
- readonly canDisable: true;
1115
- readonly outputsThinking: false;
1116
- readonly outputsSignatures: false;
1117
- };
1118
- readonly provider: "openai-responses";
1119
- } | {
1120
- readonly type: "text";
1121
- readonly modelName: "gpt-5.4-mini";
1122
- readonly description: "GPT-5.4 mini is the strongest mini model for coding, computer use, and subagents. 400K context window. Knowledge cutoff: August 2025.";
1123
- readonly maxInputTokens: 400000;
1124
- readonly maxOutputTokens: 128000;
1125
- readonly inputTokenCost: 0.75;
1126
- readonly cachedInputTokenCost: 0.075;
1127
- readonly outputTokenCost: 4.5;
1128
- readonly outputTokensPerSecond: 180;
1129
- readonly reasoning: {
1130
- readonly levels: readonly ["none", "low", "medium", "high", "xhigh"];
1131
- readonly defaultLevel: "none";
1132
- readonly canDisable: true;
1133
- readonly outputsThinking: false;
1134
- readonly outputsSignatures: false;
1135
- };
1136
- readonly provider: "openai-responses";
1137
- } | {
1138
- readonly type: "text";
1139
- readonly modelName: "gpt-5.4-nano";
1140
- readonly description: "GPT-5.4 nano is the fastest and most affordable GPT-5.4 variant for classification, data extraction, ranking, and sub-agents. 400K context window. Knowledge cutoff: August 2025.";
1141
- readonly maxInputTokens: 400000;
1142
- readonly maxOutputTokens: 128000;
1143
- readonly inputTokenCost: 0.2;
1144
- readonly cachedInputTokenCost: 0.02;
1145
- readonly outputTokenCost: 1.25;
1146
- readonly outputTokensPerSecond: 160;
1147
- readonly reasoning: {
1148
- readonly levels: readonly ["none", "low", "medium", "high", "xhigh"];
1149
- readonly defaultLevel: "none";
1150
- readonly canDisable: true;
1151
- readonly outputsThinking: false;
1152
- readonly outputsSignatures: false;
1153
- };
1154
- readonly provider: "openai-responses";
1155
- } | {
1156
- readonly type: "text";
1157
- readonly modelName: "gpt-5.4-pro";
1158
- readonly description: "GPT-5.4 Pro uses more compute for complex reasoning tasks. 1M context window. Standard pricing for ≤272K tokens. Knowledge cutoff: August 2025.";
1159
- readonly maxInputTokens: 1050000;
1160
- readonly maxOutputTokens: 128000;
1161
- readonly inputTokenCost: 30;
1162
- readonly outputTokenCost: 180;
1163
- readonly reasoning: {
1164
- readonly levels: readonly ["medium", "high", "xhigh"];
1165
- readonly defaultLevel: "medium";
1166
- readonly canDisable: false;
1167
- readonly outputsThinking: false;
1168
- readonly outputsSignatures: false;
1169
- };
1170
- readonly provider: "openai-responses";
1171
- } | {
1172
- readonly type: "text";
1173
- readonly modelName: "gpt-5.5";
1174
- readonly description: "GPT-5.5 is a new class of intelligence for coding and professional work. 1M context window. Standard pricing for ≤272K tokens, 2x input/1.5x output for >272K. Knowledge cutoff: December 2025.";
1175
- readonly maxInputTokens: 1050000;
1176
- readonly maxOutputTokens: 128000;
1177
- readonly inputTokenCost: 5;
1178
- readonly cachedInputTokenCost: 0.5;
1179
- readonly outputTokenCost: 30;
1180
- readonly outputTokensPerSecond: 66;
1181
- readonly reasoning: {
1182
- readonly levels: readonly ["none", "low", "medium", "high", "xhigh"];
1183
- readonly defaultLevel: "medium";
1184
- readonly canDisable: true;
1185
- readonly outputsThinking: false;
1186
- readonly outputsSignatures: false;
1187
- };
1188
- readonly provider: "openai-responses";
1189
- } | {
1190
- readonly type: "text";
1191
- readonly modelName: "gpt-5.5-pro";
1192
- readonly description: "GPT-5.5 Pro uses more compute for complex reasoning tasks. 1M context window. Knowledge cutoff: December 2025.";
1193
- readonly maxInputTokens: 1050000;
1194
- readonly maxOutputTokens: 128000;
1195
- readonly inputTokenCost: 30;
1196
- readonly outputTokenCost: 180;
1197
- readonly reasoning: {
1198
- readonly levels: readonly ["none", "low", "medium", "high", "xhigh"];
1199
- readonly defaultLevel: "medium";
1200
- readonly canDisable: true;
1201
- readonly outputsThinking: false;
1202
- readonly outputsSignatures: false;
1203
- };
1204
- readonly provider: "openai-responses";
1205
- } | {
1206
- readonly type: "text";
1207
- readonly modelName: "gemini-3.1-pro-preview";
1208
- readonly description: "Latest Gemini 3.1 Pro with 1M context window and 64K output. Standard pricing for ≤200k tokens ($2.00 input/$12.00 output), higher rates for >200k tokens ($4.00 input/$18.00 output). Released Feb 2026.";
1209
- readonly maxInputTokens: 1048576;
1210
- readonly maxOutputTokens: 65536;
1211
- readonly inputTokenCost: 2;
1212
- readonly outputTokenCost: 12;
1213
- readonly outputTokensPerSecond: 133;
1214
- readonly reasoning: {
1215
- readonly levels: readonly ["low", "medium", "high"];
1216
- readonly defaultLevel: "high";
1217
- readonly canDisable: false;
1218
- readonly outputsThinking: true;
1219
- readonly outputsSignatures: true;
1220
- };
1221
- readonly provider: "google";
1222
- } | {
1223
- readonly type: "text";
1224
- readonly modelName: "gemini-3-pro-preview";
1225
- readonly description: "DEPRECATED: Shut down March 9, 2026. Use gemini-3.1-pro-preview instead.";
1226
- readonly maxInputTokens: 1048576;
1227
- readonly maxOutputTokens: 65536;
1228
- readonly inputTokenCost: 2;
1229
- readonly outputTokenCost: 12;
1230
- readonly disabled: true;
1231
- readonly provider: "google";
1232
- } | {
1233
- readonly type: "text";
1234
- readonly modelName: "gemini-3.5-flash";
1235
- readonly description: "Latest Gemini 3.5 Flash model (GA May 2026). Outperforms Gemini 3.1 Pro on coding and agentic suites at 4x the speed. 1M context window, 64K output. Context caching: $0.15/1M read.";
1236
- readonly maxInputTokens: 1048576;
1237
- readonly maxOutputTokens: 65536;
1238
- readonly inputTokenCost: 1.5;
1239
- readonly cachedInputTokenCost: 0.15;
1240
- readonly outputTokenCost: 9;
1241
- readonly reasoning: {
1242
- readonly levels: readonly ["minimal", "low", "medium", "high"];
1243
- readonly defaultLevel: "high";
1244
- readonly canDisable: false;
1245
- readonly outputsThinking: true;
1246
- readonly outputsSignatures: true;
1247
- };
1248
- readonly provider: "google";
1249
- } | {
1250
- readonly type: "text";
1251
- readonly modelName: "gemini-3-flash-preview";
1252
- readonly description: "Gemini 3 Flash preview. Superseded by gemini-3.5-flash. 1M context window and 64K output. Optimized for agentic workflows and coding.";
1253
- readonly maxInputTokens: 1048576;
1254
- readonly maxOutputTokens: 65536;
1255
- readonly inputTokenCost: 0.5;
1256
- readonly outputTokenCost: 3;
1257
- readonly outputTokensPerSecond: 184;
1258
- readonly reasoning: {
1259
- readonly levels: readonly ["minimal", "low", "medium", "high"];
1260
- readonly defaultLevel: "high";
1261
- readonly canDisable: false;
1262
- readonly outputsThinking: true;
1263
- readonly outputsSignatures: true;
1264
- };
1265
- readonly provider: "google";
1266
- } | {
1267
- readonly type: "text";
1268
- readonly modelName: "gemini-3.1-flash-lite";
1269
- readonly description: "Most cost-effective Gemini 3.1 model (GA). Thinking support, 1M context window, 64K output. 2.5x faster TTFA and 45% faster output than 2.5 Flash.";
1270
- readonly maxInputTokens: 1048576;
1271
- readonly maxOutputTokens: 65536;
1272
- readonly inputTokenCost: 0.25;
1273
- readonly outputTokenCost: 1.5;
1274
- readonly outputTokensPerSecond: 379;
1275
- readonly reasoning: {
1276
- readonly levels: readonly ["minimal", "low", "medium", "high"];
1277
- readonly defaultLevel: "minimal";
1278
- readonly canDisable: false;
1279
- readonly outputsThinking: true;
1280
- readonly outputsSignatures: true;
1281
- };
1282
- readonly provider: "google";
1283
- } | {
1284
- readonly type: "text";
1285
- readonly modelName: "gemini-3.1-flash-lite-preview";
1286
- readonly description: "DEPRECATED: Preview version, discontinued July 9, 2026. Use gemini-3.1-flash-lite instead.";
1287
- readonly maxInputTokens: 1048576;
1288
- readonly maxOutputTokens: 65536;
1289
- readonly inputTokenCost: 0.25;
1290
- readonly outputTokenCost: 1.5;
1291
- readonly disabled: true;
1292
- readonly provider: "google";
1293
- } | {
1294
- readonly type: "text";
1295
- readonly modelName: "gemini-2.5-pro";
1296
- readonly description: "High-performance Gemini 2.5 model with 2M context window. Adaptive thinking for complex reasoning and coding. Standard pricing for ≤200k tokens ($1.25 input/$10.00 output), higher rates for >200k tokens ($2.50 input/$15.00 output). Batch API: 50% discount.";
1297
- readonly maxInputTokens: 2097152;
1298
- readonly maxOutputTokens: 65536;
1299
- readonly inputTokenCost: 1.25;
1300
- readonly cachedInputTokenCost: 0.31;
1301
- readonly outputTokenCost: 10;
1302
- readonly outputTokensPerSecond: 145;
1303
- readonly reasoning: {
1304
- readonly canDisable: false;
1305
- readonly outputsThinking: true;
1306
- readonly outputsSignatures: true;
1307
- };
1308
- readonly provider: "google";
1309
- } | {
1310
- readonly type: "text";
1311
- readonly modelName: "gemini-2.5-flash";
1312
- readonly description: "Balanced Gemini 2.5 model with excellent performance-to-cost ratio. Lightning-fast with controllable thinking budgets. 1M context window. Context caching available for up to 75% cost reduction.";
1313
- readonly maxInputTokens: 1048576;
1314
- readonly maxOutputTokens: 65536;
1315
- readonly inputTokenCost: 0.3;
1316
- readonly cachedInputTokenCost: 0.075;
1317
- readonly outputTokenCost: 2.5;
1318
- readonly outputTokensPerSecond: 245;
1319
- readonly reasoning: {
1320
- readonly canDisable: true;
1321
- readonly outputsThinking: true;
1322
- readonly outputsSignatures: true;
1323
- };
1324
- readonly provider: "google";
1325
- } | {
1326
- readonly type: "text";
1327
- readonly modelName: "gemini-2.5-flash-lite";
1328
- readonly description: "Most cost-effective Gemini 2.5 option for high-throughput applications. 1M context window.";
1329
- readonly maxInputTokens: 1048576;
1330
- readonly maxOutputTokens: 65536;
1331
- readonly inputTokenCost: 0.1;
1332
- readonly cachedInputTokenCost: 0.025;
1333
- readonly outputTokenCost: 0.4;
1334
- readonly outputTokensPerSecond: 400;
1335
- readonly reasoning: {
1336
- readonly canDisable: true;
1337
- readonly outputsThinking: true;
1338
- readonly outputsSignatures: false;
1339
- };
1340
- readonly provider: "google";
1341
- } | {
1342
- readonly type: "text";
1343
- readonly modelName: "gemini-2.0-flash";
1344
- readonly description: "Workhorse model for all daily tasks. Strong overall performance and supports real-time streaming Live API. 1M context window. DEPRECATED: Will be shut down on March 31, 2026.";
1345
- readonly maxInputTokens: 1048576;
1346
- readonly maxOutputTokens: 8192;
1347
- readonly inputTokenCost: 0.1;
1348
- readonly outputTokenCost: 0.4;
1349
- readonly outputTokensPerSecond: 213;
1350
- readonly disabled: true;
1351
- readonly provider: "google";
1352
- } | {
1353
- readonly type: "text";
1354
- readonly modelName: "gemini-2.0-pro-exp-02-05";
1355
- readonly description: "Strongest model quality, especially for code & world knowledge; 2M long context. In private beta.";
1356
- readonly maxInputTokens: 2097152;
1357
- readonly maxOutputTokens: 8192;
1358
- readonly inputTokenCost: 0.5;
1359
- readonly outputTokenCost: 1.5;
1360
- readonly disabled: true;
1361
- readonly provider: "google";
1362
- } | {
1363
- readonly type: "text";
1364
- readonly modelName: "gemini-2.0-flash-lite";
1365
- readonly description: "Cost effective offering to support high throughput. DEPRECATED: Will be shut down on March 31, 2026. Use gemini-2.5-flash-lite instead.";
1366
- readonly maxInputTokens: 1048576;
1367
- readonly maxOutputTokens: 8192;
1368
- readonly inputTokenCost: 0.075;
1369
- readonly outputTokenCost: 0.3;
1370
- readonly disabled: true;
1371
- readonly provider: "google";
1372
- } | {
1373
- readonly type: "text";
1374
- readonly modelName: "gemini-1.5-flash";
1375
- readonly description: "RETIRED: No longer available. Use gemini-2.5-flash instead.";
1376
- readonly maxInputTokens: 1048576;
1377
- readonly maxOutputTokens: 8192;
1378
- readonly inputTokenCost: 0.01875;
1379
- readonly outputTokenCost: 0.075;
1380
- readonly outputTokensPerSecond: 178;
1381
- readonly costUnit: "characters";
1382
- readonly disabled: true;
1383
- readonly provider: "google";
1384
- } | {
1385
- readonly type: "text";
1386
- readonly modelName: "gemini-1.5-pro";
1387
- readonly description: "RETIRED: No longer available. Use gemini-2.5-pro instead.";
1388
- readonly maxInputTokens: 2097152;
1389
- readonly maxOutputTokens: 8192;
1390
- readonly inputTokenCost: 0.3125;
1391
- readonly outputTokenCost: 1.25;
1392
- readonly outputTokensPerSecond: 59;
1393
- readonly costUnit: "characters";
1394
- readonly disabled: true;
1395
- readonly provider: "google";
1396
- } | {
1397
- readonly type: "text";
1398
- readonly modelName: "gemini-1.0-pro";
1399
- readonly description: "RETIRED: No longer available. Use gemini-2.5-flash instead.";
1400
- readonly maxInputTokens: 32760;
1401
- readonly maxOutputTokens: 8192;
1402
- readonly inputTokenCost: 0.125;
1403
- readonly outputTokenCost: 0.375;
1404
- readonly costUnit: "characters";
1405
- readonly disabled: true;
1406
- readonly provider: "google";
1407
- } | {
1408
- readonly type: "text";
1409
- readonly modelName: "claude-opus-4-8";
1410
- readonly description: "The most capable Claude model for complex reasoning and agentic coding. Same per-token pricing as Opus 4.7 with improved tool-use efficiency (~290 tokens for tool-use system prompt vs 675 on 4.7). 1M context window, 128K max output.";
1411
- readonly maxInputTokens: 1000000;
1412
- readonly maxOutputTokens: 128000;
1413
- readonly inputTokenCost: 5;
1414
- readonly cachedInputTokenCost: 0.5;
1415
- readonly outputTokenCost: 25;
1416
- readonly reasoning: {
1417
- readonly thinkingStyle: "adaptive";
1418
- readonly canDisable: false;
1419
- readonly outputsThinking: true;
1420
- readonly outputsSignatures: true;
1421
- };
1422
- readonly provider: "anthropic";
1423
- } | {
1424
- readonly type: "text";
1425
- readonly modelName: "claude-opus-4-7";
1426
- readonly description: "Claude Opus 4.7 for complex reasoning and agentic coding. Features Adaptive Thinking that auto-tunes reasoning compute per request. 1M context window, 128K max output. Knowledge cutoff: January 2026.";
1427
- readonly maxInputTokens: 1000000;
1428
- readonly maxOutputTokens: 128000;
1429
- readonly inputTokenCost: 5;
1430
- readonly cachedInputTokenCost: 0.5;
1431
- readonly cacheCreationInputTokenCost: 6.25;
1432
- readonly outputTokenCost: 25;
1433
- readonly outputTokensPerSecond: 72;
1434
- readonly reasoning: {
1435
- readonly thinkingStyle: "adaptive";
1436
- readonly canDisable: false;
1437
- readonly outputsThinking: true;
1438
- readonly outputsSignatures: true;
1439
- };
1440
- readonly provider: "anthropic";
1441
- } | {
1442
- readonly type: "text";
1443
- readonly modelName: "claude-opus-4-6";
1444
- readonly description: "Claude Opus 4.6 — legacy model. Use claude-opus-4-7 instead. 1M context window, 128K max output.";
1445
- readonly maxInputTokens: 1000000;
1446
- readonly maxOutputTokens: 128000;
1447
- readonly inputTokenCost: 5;
1448
- readonly cachedInputTokenCost: 0.5;
1449
- readonly cacheCreationInputTokenCost: 6.25;
1450
- readonly outputTokenCost: 25;
1451
- readonly outputTokensPerSecond: 53;
1452
- readonly reasoning: {
1453
- readonly thinkingStyle: "adaptive";
1454
- readonly canDisable: true;
1455
- readonly outputsThinking: true;
1456
- readonly outputsSignatures: true;
1457
- };
1458
- readonly provider: "anthropic";
1459
- } | {
1460
- readonly type: "text";
1461
- readonly modelName: "claude-sonnet-4-6";
1462
- readonly description: "The best combination of speed and intelligence. 1M context window, 64K max output.";
1463
- readonly maxInputTokens: 1000000;
1464
- readonly maxOutputTokens: 64000;
1465
- readonly inputTokenCost: 3;
1466
- readonly cachedInputTokenCost: 0.3;
1467
- readonly cacheCreationInputTokenCost: 3.75;
1468
- readonly outputTokenCost: 15;
1469
- readonly outputTokensPerSecond: 52;
1470
- readonly reasoning: {
1471
- readonly thinkingStyle: "adaptive";
1472
- readonly canDisable: true;
1473
- readonly outputsThinking: true;
1474
- readonly outputsSignatures: true;
1475
- };
1476
- readonly provider: "anthropic";
1477
- } | {
1478
- readonly type: "text";
1479
- readonly modelName: "claude-haiku-4-5-20251001";
1480
- readonly description: "The fastest Claude model with near-frontier intelligence. 200K context window, 64K max output.";
1481
- readonly maxInputTokens: 200000;
1482
- readonly maxOutputTokens: 64000;
1483
- readonly inputTokenCost: 1;
1484
- readonly cachedInputTokenCost: 0.1;
1485
- readonly cacheCreationInputTokenCost: 1.25;
1486
- readonly outputTokenCost: 5;
1487
- readonly outputTokensPerSecond: 97;
1488
- readonly reasoning: {
1489
- readonly thinkingStyle: "budget";
1490
- readonly canDisable: true;
1491
- readonly outputsThinking: true;
1492
- readonly outputsSignatures: true;
1493
- };
1494
- readonly provider: "anthropic";
1495
- } | {
1496
- readonly type: "text";
1497
- readonly modelName: "claude-3-7-sonnet-latest";
1498
- readonly description: "Claude 3.7 Sonnet — legacy model. Use claude-sonnet-4-6 instead.";
1499
- readonly maxInputTokens: 200000;
1500
- readonly maxOutputTokens: 8192;
1501
- readonly inputTokenCost: 3;
1502
- readonly outputTokenCost: 15;
1503
- readonly outputTokensPerSecond: 78;
1504
- readonly reasoning: {
1505
- readonly thinkingStyle: "budget";
1506
- readonly canDisable: true;
1507
- readonly outputsThinking: true;
1508
- readonly outputsSignatures: true;
1509
- };
1510
- readonly disabled: true;
1511
- readonly provider: "anthropic";
1512
- } | {
1513
- readonly type: "text";
1514
- readonly modelName: "claude-3-5-haiku-latest";
1515
- readonly description: "Claude 3.5 Haiku — legacy model. Use claude-haiku-4-5-20251001 instead.";
1516
- readonly maxInputTokens: 200000;
1517
- readonly maxOutputTokens: 8192;
1518
- readonly inputTokenCost: 0.8;
1519
- readonly outputTokenCost: 4;
1520
- readonly outputTokensPerSecond: 66;
1521
- readonly disabled: true;
1522
- readonly provider: "anthropic";
1523
- } | {
1524
- readonly type: "image";
1525
- readonly modelName: "gpt-image-1";
1526
- readonly provider: "openai";
1527
- readonly inputTokenCost: 5;
1528
- readonly cachedInputTokenCost: 1.25;
1529
- readonly inputImageTokenCost: 10;
1530
- readonly outputImageTokenCost: 40;
1531
- readonly costPerImage: 0.25;
1532
- } | {
1533
- readonly type: "image";
1534
- readonly modelName: "gemini-2.5-flash-image";
1535
- readonly provider: "google";
1536
- readonly description: "aka nano-banana. Graduated from preview (was gemini-2.5-flash-image-preview).";
1537
- readonly costPerImage: 0.04;
1538
- } | {
1539
- readonly type: "image";
1540
- readonly modelName: "nano-banana-pro-preview";
1541
- readonly provider: "google";
1542
- readonly description: "Higher-tier nano-banana. Preview model.";
1543
- readonly costPerImage: 0.04;
1544
- } | {
1545
- readonly type: "image";
1546
- readonly modelName: "gemini-3-pro-image-preview";
1547
- readonly provider: "google";
1548
- readonly description: "High-fidelity image generation with reasoning-enhanced composition. Supports legible text rendering, complex multi-turn editing, and character consistency using up to 14 reference inputs.";
1549
- readonly costPerImage: 0.05;
1550
- } | {
1551
- readonly type: "image";
1552
- readonly modelName: "gemini-3.1-flash-image-preview";
1553
- readonly provider: "google";
1554
- readonly description: "DEPRECATED: Preview version. Use gemini-3.1-flash-image instead.";
1555
- readonly costPerImage: 0.067;
1556
- readonly disabled: true;
1557
- } | {
1558
- readonly type: "image";
1559
- readonly modelName: "gemini-3.1-flash-image";
1560
- readonly provider: "google";
1561
- readonly description: "Fast image generation with Gemini 3.1 Flash (GA). Supports resolutions from 512px to 4096px. ~$0.045/image at 512px, $0.067 at 1K, $0.101 at 2K, $0.151 at 4K.";
1562
- readonly costPerImage: 0.067;
1563
- } | undefined;
1450
+ export declare function registerModelData(blob: ModelDataBlob): void;
1451
+ export declare function clearModelData(): void;
1452
+ export declare function getRegisteredModelData(): ModelDataBlob | null;
1453
+ export declare function getModel(modelName: ModelName, requestData?: ModelDataBlob): ModelType | undefined;
1454
+ export declare function getHostedTools(opts?: {
1455
+ provider?: string;
1456
+ model?: string;
1457
+ category?: string;
1458
+ includeDisabled?: boolean;
1459
+ modelData?: ModelDataBlob;
1460
+ }): HostedTool[];
1461
+ export declare function hostedToolPricingFor(tool: HostedTool, model?: string): HostedToolPrice | undefined;
1564
1462
  export declare function isImageModel(model: ModelType): model is ImageModel;
1565
1463
  export declare function isTextModel(model: ModelType): model is TextModel;
1566
1464
  export declare function isSpeechToTextModel(model: ModelType): model is SpeechToTextModel;