@grida/ai-models 1.20261110.1 → 1.20261110.2

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/PROVENANCE.md CHANGED
@@ -1574,3 +1574,51 @@ Adaptive thinking remains always on, with high as the default effort; sampling,
1574
1574
  assistant-prefill and forced tool selection are outside the advertised contract.
1575
1575
  The public reference discloses the model's 30-day retention requirement.
1576
1576
  Native model facts and the other provider bindings remain unchanged.
1577
+
1578
+ ## `"google/gemini-4-carbon": {`
1579
+
1580
+ Google's November 10, 2026 changelog establishes broad GA for the exact
1581
+ `gemini-4-carbon` model. The October 28 announcement describes restricted
1582
+ Fairwind access, which is not the catalogue's broad-public release date.
1583
+ https://aistudio.google.com/docs/changelog
1584
+ https://blog.google/innovation-and-ai/models-and-research/gemini-models/introducing-gemini-4-carbon/
1585
+
1586
+ Regular Standard global rates are independently published as $4 input,
1587
+ $20 output (including thinking) and $0.20 cached input per million tokens.
1588
+ Google states those regular values directly; Vercel's model page separately
1589
+ publishes `inputListCostTiers`, `outputListCostTiers` and
1590
+ `cachedInputListCostTiers` with `priceType: "list"`, amounts 4, 20 and 0.2,
1591
+ and a single unbounded band starting at zero for each meter. OpenRouter's
1592
+ provider rows explicitly show the same regular rates alongside discounted
1593
+ prices. No regular price is inferred by reversing a promotion. The card
1594
+ excludes introductory prices, alternate service tiers, regional premiums
1595
+ and optional grounding charges. Implicit caching has no sourced per-token
1596
+ write charge; the unresolved OpenRouter write field remains qualified.
1597
+ https://aistudio.google.com/docs/pricing
1598
+ https://vercel.com/ai-gateway/models/gemini-4-carbon
1599
+ https://openrouter.ai/google/gemini-4-carbon
1600
+ https://openrouter.ai/api/v1/models
1601
+
1602
+ The unified window is 1,048,576 input, thinking and output tokens. Native
1603
+ foreground hops permit 65,536 tokens with streaming and 32,768 without it.
1604
+ Long decode uses background execution or continuation, reprocessing earlier
1605
+ output as input on later hops; the reported 1M provider ceiling is not a
1606
+ single synchronous-call guarantee. The intrinsic/native output field
1607
+ records the 65,536-token streaming-hop limit, while provider bindings
1608
+ preserve their own discovery limits and these qualifications.
1609
+ https://aistudio.google.com/docs/models/gemini-4-carbon
1610
+ https://aistudio.google.com/docs/long-decode
1611
+ https://ai-gateway.vercel.sh/v1/models
1612
+
1613
+ Thinking supports low, medium and high, defaulting to medium. Sampling
1614
+ controls are ignored. Eligible prompts use implicit caching; explicit
1615
+ cache creation is unsupported. Text, images, video, audio and PDF input
1616
+ produce text; Vercel's listed modalities are independently narrower.
1617
+ The Gemini 4/3 media-resolution table supplies approximate image budgets
1618
+ of 280/560/1120/2240 tokens, with 1120 as the unspecified default. The image
1619
+ formats come from Google's image-understanding guide. Paid-service data
1620
+ terms do not establish zero retention.
1621
+ https://aistudio.google.com/docs/whats-new-gemini-4-carbon
1622
+ https://aistudio.google.com/docs/media-resolution
1623
+ https://aistudio.google.com/docs/image-understanding
1624
+ https://aistudio.google.com/docs/terms
package/README.md CHANGED
@@ -5,7 +5,7 @@ provider bindings, and published prices, with the labs that make them and their
5
5
  logos. Every fact cites a first-party source.
6
6
 
7
7
  This package is the data, published unchanged from release
8
- [`v1.20261110.1`](https://github.com/gridaco/models/releases/tag/v1.20261110.1) of [gridaco/models](https://github.com/gridaco/models). It contains no code.
8
+ [`v1.20261110.2`](https://github.com/gridaco/models/releases/tag/v1.20261110.2) of [gridaco/models](https://github.com/gridaco/models). It contains no code.
9
9
 
10
10
  | file | contents |
11
11
  | --- | --- |
@@ -25,7 +25,7 @@ From a CDN, without installing:
25
25
 
26
26
  ```
27
27
  https://cdn.jsdelivr.net/npm/@grida/ai-models@1/models.json
28
- https://cdn.jsdelivr.net/npm/@grida/ai-models@1.20261110.1/models.json
28
+ https://cdn.jsdelivr.net/npm/@grida/ai-models@1.20261110.2/models.json
29
29
  ```
30
30
 
31
31
  ## Versions
package/manifest.json CHANGED
@@ -1,19 +1,19 @@
1
1
  {
2
2
  "name": "grida-models",
3
- "version": "v1.20261110.1",
3
+ "version": "v1.20261110.2",
4
4
  "schema_version": 1,
5
- "commit": "f895d86fd5db2ec1af5654de89016aee36783f56",
6
- "created": "2026-11-10T12:01:08Z",
7
- "source": "https://github.com/gridaco/models/tree/f895d86fd5db2ec1af5654de89016aee36783f56/catalogue",
8
- "digest": "sha256:15d78d245eca4a73367c6b6442b1a5363d2fb8c31ced85362af3834c4b8fe9aa",
5
+ "commit": "6f470a8f3d67642203f7dae1c619a5a4b96d6ca7",
6
+ "created": "2026-11-10T17:26:55Z",
7
+ "source": "https://github.com/gridaco/models/tree/6f470a8f3d67642203f7dae1c619a5a4b96d6ca7/catalogue",
8
+ "digest": "sha256:0f4989b306b5a9e00ce4619a5c64d066f8f28834ae161232b09867b1fd20b2d8",
9
9
  "counts": {
10
- "cards": 80,
10
+ "cards": 81,
11
11
  "labs": 18
12
12
  },
13
13
  "files": {
14
14
  "PROVENANCE.md": {
15
- "sha256": "f5fb9dfb966b4da8f2497fbe4c9bd34eec5260a6d047c8d3b2a2aa1d79a422e5",
16
- "bytes": 79476
15
+ "sha256": "e5b9a201c3f10e40b7d30f6df33371477b15895a115da5b1751648548822c5cd",
16
+ "bytes": 82229
17
17
  },
18
18
  "README.md": {
19
19
  "sha256": "196afd4ca14e909b5a3adc6bb7bc78f823c45f29163b22047dbc5c30e6448998",
@@ -92,8 +92,8 @@
92
92
  "bytes": 330
93
93
  },
94
94
  "models.json": {
95
- "sha256": "775dbfe4cac94acc689d8ea0a9ade7de52638f62c53a000566d55569e7b54d23",
96
- "bytes": 811583
95
+ "sha256": "cb7461d6d576fdfd36c5892ac8dd87d58872d1abf0393a4b56ebfc761bad8ce1",
96
+ "bytes": 820413
97
97
  },
98
98
  "schemas/labs.schema.json": {
99
99
  "sha256": "9ddaedd56826bebdbe72b84f76cfc86df0b5c5902ba9f0259c2ea2cbb9327f99",
package/models.json CHANGED
@@ -3338,6 +3338,199 @@
3338
3338
  }
3339
3339
  }
3340
3340
  },
3341
+ "google/gemini-4-carbon": {
3342
+ "id": "google/gemini-4-carbon",
3343
+ "vendor": "google",
3344
+ "reference_provider": "google",
3345
+ "label": "Gemini 4 Carbon",
3346
+ "multimodal": true,
3347
+ "image_input_mimes": [
3348
+ "image/png",
3349
+ "image/jpeg",
3350
+ "image/webp",
3351
+ "image/heic",
3352
+ "image/heif"
3353
+ ],
3354
+ "tool_call": true,
3355
+ "context_window": 1048576,
3356
+ "output_limit": 65536,
3357
+ "cost": {
3358
+ "input": 4,
3359
+ "output": 20,
3360
+ "cache_read": 0.2
3361
+ },
3362
+ "release": {
3363
+ "date": "2026-11-10",
3364
+ "basis": "model",
3365
+ "source_url": "https://aistudio.google.com/docs/changelog"
3366
+ },
3367
+ "vision_tokenization": {
3368
+ "source_url": "https://aistudio.google.com/docs/media-resolution",
3369
+ "maximum_tokens_per_image": 2240,
3370
+ "token_counts_approximate": true,
3371
+ "detail_token_budgets": {
3372
+ "unspecified": 1120,
3373
+ "low": 280,
3374
+ "medium": 560,
3375
+ "high": 1120,
3376
+ "ultra_high": 2240
3377
+ }
3378
+ },
3379
+ "providers": {
3380
+ "google": {
3381
+ "provider": "google",
3382
+ "id": "gemini-4-carbon",
3383
+ "url": "https://aistudio.google.com/docs/models/gemini-4-carbon",
3384
+ "source_url": "https://aistudio.google.com/docs/models/gemini-4-carbon",
3385
+ "endpoints": [
3386
+ {
3387
+ "id": "gemini-4-carbon",
3388
+ "input": "text",
3389
+ "source_url": "https://aistudio.google.com/docs/models/gemini-4-carbon"
3390
+ }
3391
+ ],
3392
+ "qualifications": [
3393
+ {
3394
+ "source_url": "https://aistudio.google.com/docs/long-decode",
3395
+ "notes": "Native foreground generation permits 65,536 tokens per streaming hop and 32,768 per non-streaming hop. Longer responses require background execution or continuation within the unified 1,048,576-token input/thinking/output window; a discovery ceiling does not establish a single synchronous hop."
3396
+ },
3397
+ {
3398
+ "source_url": "https://aistudio.google.com/docs/whats-new-gemini-4-carbon",
3399
+ "notes": "Text, image, video, audio and PDF inputs produce text. Thinking levels are low, medium and high, with medium as the default. Implicit caching is enabled for eligible prompts of at least 4,096 tokens; explicit caching is unsupported. Sampling parameters temperature, top_p and top_k are ignored."
3400
+ },
3401
+ {
3402
+ "source_url": "https://aistudio.google.com/docs/terms",
3403
+ "notes": "Paid-service prompts and responses are not used to improve Google's products, but limited logging for safety and other stated purposes remains. This does not establish zero retention; storage and logging depend on the API and applicable settings."
3404
+ }
3405
+ ],
3406
+ "limits": {
3407
+ "source_url": "https://aistudio.google.com/docs/models/gemini-4-carbon",
3408
+ "context_tokens": 1048576,
3409
+ "max_output_tokens": 65536
3410
+ },
3411
+ "capabilities": {
3412
+ "source_url": "https://aistudio.google.com/docs/models/gemini-4-carbon",
3413
+ "image_input": true,
3414
+ "tools": true,
3415
+ "reasoning": true,
3416
+ "structured_output": true
3417
+ },
3418
+ "token_pricing": {
3419
+ "source_url": "https://aistudio.google.com/docs/pricing",
3420
+ "unit": "usd_per_token",
3421
+ "scope": "Regular Standard global token rates. Implicit caching only. Introductory discounts, grounding tools and alternate service tiers are excluded.",
3422
+ "input": "0.000004",
3423
+ "output": "0.00002",
3424
+ "input_cache_read": "0.0000002"
3425
+ }
3426
+ },
3427
+ "vercel": {
3428
+ "provider": "vercel",
3429
+ "id": "google/gemini-4-carbon",
3430
+ "url": "https://vercel.com/ai-gateway/models/gemini-4-carbon",
3431
+ "source_url": "https://vercel.com/ai-gateway/models/gemini-4-carbon",
3432
+ "endpoints": [
3433
+ {
3434
+ "id": "google/gemini-4-carbon",
3435
+ "input": "text",
3436
+ "source_url": "https://vercel.com/ai-gateway/models/gemini-4-carbon"
3437
+ }
3438
+ ],
3439
+ "qualifications": [
3440
+ {
3441
+ "source_url": "https://aistudio.google.com/docs/long-decode",
3442
+ "notes": "Native foreground generation permits 65,536 tokens per streaming hop and 32,768 per non-streaming hop. Longer responses require background execution or continuation within the unified input/thinking/output window; the provider's 1,000,000-token discovery ceiling does not establish a single synchronous hop."
3443
+ }
3444
+ ],
3445
+ "limits": {
3446
+ "source_url": "https://ai-gateway.vercel.sh/v1/models",
3447
+ "context_tokens": 1000000,
3448
+ "max_output_tokens": 1000000
3449
+ },
3450
+ "capabilities": {
3451
+ "source_url": "https://ai-gateway.vercel.sh/v1/models",
3452
+ "image_input": true,
3453
+ "tools": true,
3454
+ "reasoning": true,
3455
+ "structured_output": true
3456
+ },
3457
+ "token_pricing": {
3458
+ "source_url": "https://vercel.com/ai-gateway/models/gemini-4-carbon",
3459
+ "unit": "usd_per_token",
3460
+ "scope": "Regular Standard global list token rates from the provider's explicit list-price fields. Implicit caching only. Introductory discounts, grounding tools, alternate service tiers and regional premiums are excluded.",
3461
+ "input": "0.000004",
3462
+ "output": "0.00002",
3463
+ "input_cache_read": "0.0000002"
3464
+ }
3465
+ },
3466
+ "openrouter": {
3467
+ "provider": "openrouter",
3468
+ "id": "google/gemini-4-carbon",
3469
+ "url": "https://openrouter.ai/google/gemini-4-carbon",
3470
+ "source_url": "https://openrouter.ai/google/gemini-4-carbon",
3471
+ "endpoints": [
3472
+ {
3473
+ "id": "google/gemini-4-carbon",
3474
+ "input": "text",
3475
+ "source_url": "https://openrouter.ai/google/gemini-4-carbon"
3476
+ }
3477
+ ],
3478
+ "qualifications": [
3479
+ {
3480
+ "source_url": "https://openrouter.ai/api/v1/models",
3481
+ "notes": "The model page's completion ceiling differs from the live feed's top_provider.max_completion_tokens of 943,718. This binding records the feed's explicit serving limit rather than treating a rounded 1M claim as a synchronous guarantee."
3482
+ },
3483
+ {
3484
+ "source_url": "https://aistudio.google.com/docs/long-decode",
3485
+ "notes": "Native foreground generation permits 65,536 tokens per streaming hop and 32,768 per non-streaming hop. Longer responses require background execution or continuation within the unified input/thinking/output window; provider discovery ceilings do not establish a single synchronous hop."
3486
+ },
3487
+ {
3488
+ "source_url": "https://openrouter.ai/api/v1/models",
3489
+ "notes": "Discovery includes an input_cache_write amount, but its applicable caching operation, TTL and billing unit are not established for this implicit-only model. No per-token cache-write price is recorded; omission does not mean zero cost."
3490
+ }
3491
+ ],
3492
+ "limits": {
3493
+ "source_url": "https://openrouter.ai/api/v1/models",
3494
+ "context_tokens": 1048576,
3495
+ "max_output_tokens": 943718
3496
+ },
3497
+ "capabilities": {
3498
+ "source_url": "https://openrouter.ai/api/v1/models",
3499
+ "image_input": true,
3500
+ "tools": true,
3501
+ "reasoning": true,
3502
+ "structured_output": true
3503
+ },
3504
+ "token_pricing": {
3505
+ "source_url": "https://openrouter.ai/google/gemini-4-carbon",
3506
+ "unit": "usd_per_token",
3507
+ "scope": "Regular Standard route token rates explicitly shown alongside discounted prices on the provider page. Optional tools, alternate service tiers and regional premiums are excluded.",
3508
+ "input": "0.000004",
3509
+ "output": "0.00002",
3510
+ "input_cache_read": "0.0000002"
3511
+ }
3512
+ }
3513
+ },
3514
+ "provider_coverage": {
3515
+ "google": {
3516
+ "disposition": "listed",
3517
+ "source_url": "https://aistudio.google.com/docs/models/gemini-4-carbon"
3518
+ },
3519
+ "vercel": {
3520
+ "disposition": "listed",
3521
+ "source_url": "https://vercel.com/ai-gateway/models/gemini-4-carbon"
3522
+ },
3523
+ "openrouter": {
3524
+ "disposition": "listed",
3525
+ "source_url": "https://openrouter.ai/google/gemini-4-carbon"
3526
+ },
3527
+ "fal": {
3528
+ "disposition": "excluded",
3529
+ "source_url": "https://fal.ai/models",
3530
+ "reason": "The provider coverage policy includes fal for media models only; no text API compatibility is claimed."
3531
+ }
3532
+ }
3533
+ },
3341
3534
  "openai/gpt-5.5": {
3342
3535
  "id": "openai/gpt-5.5",
3343
3536
  "vendor": "openai",
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@grida/ai-models",
3
- "version": "1.20261110.1",
3
+ "version": "1.20261110.2",
4
4
  "description": "A hand-curated, source-backed catalogue of AI models, their labs, and published prices.",
5
5
  "keywords": [
6
6
  "ai",