{
  "schema_version": 1,
  "publication": "Request Budget",
  "providers": [
    {
      "id": "groq",
      "name": "Groq",
      "mark": "GQ",
      "type": "host",
      "claims": [
        "IA-001",
        "IA-002",
        "IA-003",
        "IA-004"
      ]
    },
    {
      "id": "fireworks-ai",
      "name": "Fireworks AI",
      "mark": "FW",
      "type": "host",
      "claims": [
        "IA-005",
        "IA-006",
        "IA-007",
        "IA-008"
      ]
    },
    {
      "id": "deepinfra",
      "name": "DeepInfra",
      "mark": "DI",
      "type": "host",
      "claims": [
        "IA-009"
      ]
    },
    {
      "id": "together-ai",
      "name": "Together AI",
      "mark": "TA",
      "type": "host",
      "claims": [
        "IA-010",
        "IA-011",
        "IA-012",
        "IA-013",
        "IA-014"
      ]
    },
    {
      "id": "openai",
      "name": "OpenAI",
      "mark": "OA",
      "type": "firstparty",
      "claims": [
        "IA-015",
        "IA-016",
        "IA-017",
        "IA-018",
        "IA-041"
      ]
    },
    {
      "id": "anthropic",
      "name": "Anthropic",
      "mark": "AN",
      "type": "firstparty",
      "claims": [
        "IA-019",
        "IA-020",
        "IA-021",
        "IA-022"
      ]
    },
    {
      "id": "google-vertex-ai",
      "name": "Google Vertex AI",
      "mark": "GV",
      "type": "cloud",
      "claims": [
        "IA-023",
        "IA-024",
        "IA-025",
        "IA-026"
      ]
    },
    {
      "id": "amazon-bedrock",
      "name": "Amazon Bedrock",
      "mark": "AB",
      "type": "cloud",
      "claims": [
        "IA-027",
        "IA-028",
        "IA-029"
      ]
    },
    {
      "id": "microsoft-azure-ai-foundry",
      "name": "Microsoft Foundry",
      "mark": "MF",
      "type": "cloud",
      "claims": [
        "IA-030",
        "IA-031",
        "IA-032"
      ]
    },
    {
      "id": "openrouter",
      "name": "OpenRouter",
      "mark": "OR",
      "type": "gateway",
      "claims": [
        "IA-033",
        "IA-034",
        "IA-035"
      ]
    },
    {
      "id": "replicate",
      "name": "Replicate",
      "mark": "RE",
      "type": "host",
      "claims": [
        "IA-036"
      ]
    },
    {
      "id": "cerebras",
      "name": "Cerebras",
      "mark": "CE",
      "type": "host",
      "claims": [
        "IA-037",
        "IA-038",
        "IA-039",
        "IA-040"
      ]
    }
  ],
  "records": [
    {
      "id": "IA-001",
      "vendor": "groq",
      "field": "price",
      "label": "GPT-OSS-120B",
      "value": "$0.15 input / $0.60 output per 1M tokens",
      "source_id": "groq-models",
      "source_url": "https://console.groq.com/docs/models",
      "source_role": "vendor_claim",
      "read_at": "2026-11-14T13:14:18.585Z",
      "source_sha256": "29e04ae4326f939560e3a9d0d8acbdda0d3d43ab95f88916f1266702338c5821",
      "locator": "L169–174, production model table",
      "evidence_excerpt": "$0.15 input$0.60 output",
      "scope": "openai/gpt-oss-120b, Developer plan",
      "conditions": "Listed token rates; no cache, batch, tools, taxes or retry cost included. Serving configuration and output fidelity untested.",
      "input_rate": 0.15,
      "output_rate": 0.6,
      "model": "openai/gpt-oss-120b",
      "mode": "host",
      "claim_sha256": "51f721a86a5fa1b348ca05ddba5e447d9e529d4e2469d49406eda56d4f497dc8"
    },
    {
      "id": "IA-002",
      "vendor": "groq",
      "field": "capacity",
      "label": "Developer limits",
      "value": "250K TPM / 1K RPM",
      "source_id": "groq-models",
      "source_url": "https://console.groq.com/docs/models",
      "source_role": "vendor_claim",
      "read_at": "2026-11-14T13:14:18.585Z",
      "source_sha256": "29e04ae4326f939560e3a9d0d8acbdda0d3d43ab95f88916f1266702338c5821",
      "locator": "L169–174",
      "evidence_excerpt": "250K TPM 1K RPM",
      "scope": "GPT-OSS-120B Developer plan",
      "conditions": "Rate ceilings are not reserved throughput or achieved speed. Context 131,072; max completion 65,536 tokens.",
      "claim_sha256": "77c1812d8aad0f59e350f6f54c6a7b49048882195e98cdfac0ba5c7fa9c2e1ff"
    },
    {
      "id": "IA-003",
      "vendor": "groq",
      "field": "data",
      "label": "Inference retention",
      "value": "No customer-content retention by default; exceptions apply",
      "source_id": "groq-data",
      "source_url": "https://console.groq.com/docs/your-data",
      "source_role": "vendor_claim",
      "read_at": "2026-11-14T13:22:23.484Z",
      "source_sha256": "a6e86c4777cd1e703a70cb2fe6937872afeff3a540fff524bc1f013f6a27942d",
      "locator": "L167–195",
      "evidence_excerpt": "By default, Groq does not retain customer data for inference requests.",
      "scope": "GroqCloud inference",
      "conditions": "Persistence features and reliability/abuse troubleshooting can retain content. Logs up to 30 days unless law requires longer; Data Controls opt-out. Usage metadata retained.",
      "claim_sha256": "c5e18d66d1beae6cb81eebf43a72a13520d9c592fb5f30b71d977f7ec61265b1"
    },
    {
      "id": "IA-004",
      "vendor": "groq",
      "field": "capacity",
      "label": "Model access",
      "value": "Llama 3.3 70B is listed Enterprise / Contact Sales",
      "source_id": "groq-models",
      "source_url": "https://console.groq.com/docs/models",
      "source_role": "vendor_claim",
      "read_at": "2026-11-14T13:14:18.585Z",
      "source_sha256": "29e04ae4326f939560e3a9d0d8acbdda0d3d43ab95f88916f1266702338c5821",
      "locator": "L171–179",
      "evidence_excerpt": "Enterprise llama-3.3-70b-versatile",
      "scope": "llama-3.3-70b-versatile",
      "conditions": "Do not apply old public rates to this listing. Preview models are evaluation-only and may be discontinued at short notice.",
      "claim_sha256": "4d719b87eb9dd2f948474b7ca0429200e722df922bcdff6c84b3f32e489f7119"
    },
    {
      "id": "IA-005",
      "vendor": "fireworks-ai",
      "field": "price",
      "label": "GPT-OSS-120B Standard",
      "value": "$0.15 input / $0.60 output per 1M tokens",
      "source_id": "fireworks-pricing",
      "source_url": "https://fireworks.ai/pricing",
      "source_role": "vendor_claim",
      "read_at": "2026-11-14T13:14:09.865Z",
      "source_sha256": "1d3a58dcf7e8c99c65881527493424c7c65bd0876cbe0bd20f8a738df43f80e2",
      "locator": "L103–122, Standard/priority table",
      "evidence_excerpt": "$0.15",
      "scope": "GPT-OSS-120B, Standard Serverless",
      "conditions": "Cached input $0.015 / 1M separately. Priority rates differ; US-only option on select models has a 50% premium, waived on Muse Spark Open.",
      "input_rate": 0.15,
      "output_rate": 0.6,
      "cached_rate": 0.015,
      "model": "GPT-OSS-120B",
      "mode": "host",
      "claim_sha256": "d7b6fad0f4bbf632d4bc16b3e8c3ee79ccc295954a4290185531a2b2aa678208"
    },
    {
      "id": "IA-006",
      "vendor": "fireworks-ai",
      "field": "batch",
      "label": "Batch",
      "value": "50% off input and output",
      "source_id": "fireworks-pricing",
      "source_url": "https://fireworks.ai/pricing",
      "source_role": "vendor_claim",
      "read_at": "2026-11-14T13:14:09.865Z",
      "source_sha256": "1d3a58dcf7e8c99c65881527493424c7c65bd0876cbe0bd20f8a738df43f80e2",
      "locator": "L96–101",
      "evidence_excerpt": "50% off input and output",
      "scope": "Fireworks Batch jobs",
      "conditions": "For work that does not need an immediate answer. Not a blanket claim of immediate serving at half price.",
      "claim_sha256": "893b10103ab11b255e333fa10ec6585c2dba69f3edae016e7ee017c3058fb8be"
    },
    {
      "id": "IA-007",
      "vendor": "fireworks-ai",
      "field": "capacity",
      "label": "Spot",
      "value": "30% off Standard; throughput can fall to zero",
      "source_id": "fireworks-pricing",
      "source_url": "https://fireworks.ai/pricing",
      "source_role": "vendor_claim",
      "read_at": "2026-11-14T13:14:09.865Z",
      "source_sha256": "1d3a58dcf7e8c99c65881527493424c7c65bd0876cbe0bd20f8a738df43f80e2",
      "locator": "L99–101",
      "evidence_excerpt": "Throughput can drop to zero",
      "scope": "Fireworks Spot Serverless, select models only",
      "conditions": "Retry or fall back to Standard. Not reserved capacity and not every model.",
      "claim_sha256": "a56f37f9e9967892f74eb7720b500c6f3dcb4e1891f1c7437ee7e59f4ac809b1"
    },
    {
      "id": "IA-008",
      "vendor": "fireworks-ai",
      "field": "capacity",
      "label": "Reserved Throughput",
      "value": "Separate dollars-per-minute capacity offer",
      "source_id": "fireworks-pricing",
      "source_url": "https://fireworks.ai/pricing",
      "source_role": "vendor_claim",
      "read_at": "2026-11-14T13:14:09.865Z",
      "source_sha256": "1d3a58dcf7e8c99c65881527493424c7c65bd0876cbe0bd20f8a738df43f80e2",
      "locator": "L44–56",
      "evidence_excerpt": "99.9%",
      "scope": "Fireworks Reserved Throughput",
      "conditions": "Commitment required. Vendor lists a 99.9% uptime SLA, subject to service terms. Usage above reserved capacity is charged at Standard rates; no contract remedy independently reviewed.",
      "claim_sha256": "9557d396a552d2974e4002c1b1f22c573a96ce499fd13659a2ff6656915e6a73"
    },
    {
      "id": "IA-009",
      "vendor": "deepinfra",
      "field": "price",
      "label": "GPT-OSS-120B",
      "value": "$0.037 input / $0.17 output per 1M tokens",
      "source_id": "deepinfra-gptoss",
      "source_url": "https://deepinfra.com/openai/gpt-oss-120b",
      "source_role": "vendor_claim",
      "read_at": "2026-11-14T13:14:45.928Z",
      "source_sha256": "c92a8429151c33d33085919a3522f70acf24ece488e81162c62a69182adb12d4",
      "locator": "L49–74",
      "evidence_excerpt": "$0.037",
      "scope": "openai/gpt-oss-120b, public model listing, bfloat16",
      "conditions": "131,072 context listed. Zero retention badge alone is not a reviewed endpoint policy. No independent serving-fidelity or latency test.",
      "input_rate": 0.037,
      "output_rate": 0.17,
      "model": "openai/gpt-oss-120b",
      "precision": "bfloat16",
      "mode": "host",
      "claim_sha256": "ef760d49447277acde85125d1972d5c422475d1bb32e53259769a0116cf3a127"
    },
    {
      "id": "IA-010",
      "vendor": "together-ai",
      "field": "price",
      "label": "GPT-OSS-120B",
      "value": "$0.15 input / $0.60 output per 1M tokens",
      "source_id": "together-gptoss",
      "source_url": "https://www.together.ai/models/gpt-oss-120b",
      "source_role": "vendor_claim",
      "read_at": "2026-11-14 13:32:44 UTC",
      "source_sha256": "3120d139ec354ff83df343a2ff4b9f360d64169dae8b8350e954e8138af81eba",
      "locator": "L48, L195–219",
      "evidence_excerpt": "it costs $0.15",
      "scope": "openai/gpt-oss-120b model listing; Serverless and Dedicated links",
      "conditions": "128K context listed. Listed rates, not a live account quote; page also contains a generic launching-soon component. Model availability not independently exercised.",
      "input_rate": 0.15,
      "output_rate": 0.6,
      "model": "openai/gpt-oss-120b",
      "mode": "host",
      "claim_sha256": "fbf33330224185a83561ee151ba95dc81c663d14c4f94848704e10a59b5021ef"
    },
    {
      "id": "IA-011",
      "vendor": "together-ai",
      "field": "capacity",
      "label": "Serverless",
      "value": "Best effort; 429 or 503 under demand",
      "source_id": "together-rate-limits",
      "source_url": "https://docs.together.ai/docs/serverless/rate-limits",
      "source_role": "vendor_claim",
      "read_at": "2026-11-13T04:25:07.967Z",
      "source_sha256": "0e468d21026e02e9d4715b6dceda4d45497bb3578cbdc0edc14a68bb1c806b7d",
      "locator": "text lines 8–26, Demand and performance",
      "evidence_excerpt": "Serverless performance is best-effort.",
      "scope": "Together Serverless",
      "conditions": "Limit retries to the application latency budget. Provisioned throughput reserves selected model/family capacity under a separate SLA.",
      "claim_sha256": "ae8da2f21ba5377deb832c1d0c6ab8f0bacf5d478bd994d10f1fc48f0a8ae452"
    },
    {
      "id": "IA-012",
      "vendor": "together-ai",
      "field": "data",
      "label": "Organization privacy controls",
      "value": "Prompt storage on; model training opt-in off",
      "source_id": "together-privacy",
      "source_url": "https://docs.together.ai/docs/privacy-and-security",
      "source_role": "vendor_claim",
      "read_at": "2026-11-13T04:25:08.899Z",
      "source_sha256": "0b709b5615d0112c857094e19eb78cd84c71cc32a80913476a07dd9c1ddac50b",
      "locator": "text lines 8–17, Privacy controls",
      "evidence_excerpt": "Prompt storage: On by default.",
      "scope": "Together organization privacy settings",
      "conditions": "Turning prompt storage off enables ZDR and disables passthrough models. Passthrough sends data to upstream providers under their own policy.",
      "claim_sha256": "feab38bf18253885dad87cdef04ecc2a286dc7b03dd24556cc9f5c83fe8de8f6"
    },
    {
      "id": "IA-013",
      "vendor": "together-ai",
      "field": "region",
      "label": "US processing endpoint",
      "value": "Select Serverless models can process in the US",
      "source_id": "together-privacy",
      "source_url": "https://docs.together.ai/docs/privacy-and-security",
      "source_role": "vendor_claim",
      "read_at": "2026-11-13T04:25:08.899Z",
      "source_sha256": "0b709b5615d0112c857094e19eb78cd84c71cc32a80913476a07dd9c1ddac50b",
      "locator": "US-only processing paragraph",
      "evidence_excerpt": "us-api.together.ai",
      "scope": "us-api.together.ai, select Serverless models",
      "conditions": "This processing scope does not guarantee storage, logging or routing location. Other residency requirements may require dedicated/VPC deployment.",
      "claim_sha256": "e20accb50c8b0b90f6e00087375f399deda8b0b65afbfa8f00db987b9a3167b8"
    },
    {
      "id": "IA-014",
      "vendor": "together-ai",
      "field": "features",
      "label": "Hosted-model fidelity",
      "value": "Together says it hosts at full precision",
      "source_id": "together-privacy",
      "source_url": "https://docs.together.ai/docs/privacy-and-security",
      "source_role": "vendor_claim",
      "read_at": "2026-11-13T04:25:08.899Z",
      "source_sha256": "0b709b5615d0112c857094e19eb78cd84c71cc32a80913476a07dd9c1ddac50b",
      "locator": "Third-party model providers, final paragraph",
      "evidence_excerpt": "does not distill them, force system prompts",
      "scope": "Non-passthrough models hosted on Together",
      "conditions": "Vendor statement only. No independent output-fidelity test. Do not infer identical behavior across hosts or precision from the base model name.",
      "claim_sha256": "155e91fe19d1e1e04ca2c2f814c21cd4665b11fa407792bec6ecb912d2c5cb14"
    },
    {
      "id": "IA-015",
      "vendor": "openai",
      "field": "price",
      "label": "GPT-6.1 Sol Standard",
      "value": "$2 input / $10 output per 1M tokens",
      "source_id": "openai-sol",
      "source_url": "https://developers.openai.com/api/docs/models/gpt-6.1-sol",
      "source_role": "vendor_claim",
      "read_at": "2026-11-14 13:32:44 UTC",
      "source_sha256": "39f22f9fbca0e062165e8a6c896c0d2e3a37f125b6c21abd244ec131be100016",
      "locator": "L999–1028",
      "evidence_excerpt": "$2.00",
      "scope": "gpt-6.1-sol API text tokens, Standard",
      "conditions": "Cached input $0.10; cache write $2.50. Over 272K input tokens: full request input/cache rates 2×, output 1.5×. Regional processing adds 10% where offered.",
      "input_rate": 2,
      "output_rate": 10,
      "model": "gpt-6.1-sol",
      "mode": "firstparty",
      "claim_sha256": "4c63c882a1f089d3b568c50501bd6736eeb537e832c1174be9f6d24b5c8573d6"
    },
    {
      "id": "IA-016",
      "vendor": "openai",
      "field": "features",
      "label": "Tool calling",
      "value": "Requires Responses API for GPT-6.1 Sol",
      "source_id": "openai-sol",
      "source_url": "https://developers.openai.com/api/docs/models/gpt-6.1-sol",
      "source_role": "vendor_claim",
      "read_at": "2026-11-14 13:32:44 UTC",
      "source_sha256": "39f22f9fbca0e062165e8a6c896c0d2e3a37f125b6c21abd244ec131be100016",
      "locator": "L982–989",
      "evidence_excerpt": "Tool calling requires the Responses API.",
      "scope": "gpt-6.1-sol",
      "conditions": "Context 1,050,000 and max output 128,000 tokens. Do not infer Chat Completions tool parity from an OpenAI-compatible request shape.",
      "claim_sha256": "5bf4f53f889264d23098e1bf66ecf48cb3a185fec8e810f0f476b22ec4520634"
    },
    {
      "id": "IA-017",
      "vendor": "openai",
      "field": "data",
      "label": "Default API handling",
      "value": "No training without opt-in; abuse logs up to 30 days",
      "source_id": "openai-api-data-use",
      "source_url": "https://developers.openai.com/api/docs/guides/your-data",
      "source_role": "vendor_claim",
      "read_at": "2026-11-13T04:25:05.526Z",
      "source_sha256": "17d804a75634e3dbeae598f81998deb0ac471a61710ed5b64bd594bb6d855580",
      "locator": "text lines 6–21",
      "evidence_excerpt": "retained for up to 30 days",
      "scope": "OpenAI API",
      "conditions": "Legal/safety exceptions can extend retention. ZDR requires approval; endpoint and feature limitations may still retain application state. Responses default application state is a separate control.",
      "claim_sha256": "cf35cf1240d61f1ded8f66157d603b13149a62611056e59c8e6126cdc264ef37"
    },
    {
      "id": "IA-018",
      "vendor": "openai",
      "field": "batch",
      "label": "Batch and Flex",
      "value": "50% lower than Standard",
      "source_id": "openai-sol",
      "source_url": "https://developers.openai.com/api/docs/models/gpt-6.1-sol",
      "source_role": "vendor_claim",
      "read_at": "2026-11-14 13:32:44 UTC",
      "source_sha256": "39f22f9fbca0e062165e8a6c896c0d2e3a37f125b6c21abd244ec131be100016",
      "locator": "L1022–1028",
      "evidence_excerpt": "Batch and Flex prices are 50% lower than Standard.",
      "scope": "gpt-6.1-sol API",
      "conditions": "Fast 2× Standard; Ultrafast 6×. Long-context and regional conditions still matter; subscription use has different terms.",
      "claim_sha256": "53086c1d549000fc932333f39db109eddc95c442e9ff6217c876cf5d10e25b34"
    },
    {
      "id": "IA-019",
      "vendor": "anthropic",
      "field": "price",
      "label": "Sonnet 5.5 base tokens",
      "value": "$2 input / $10 output per MTok",
      "source_id": "anthropic-pricing",
      "source_url": "https://platform.claude.com/docs/en/about-claude/pricing",
      "source_role": "vendor_claim",
      "read_at": "2026-11-14T13:15:04.957Z",
      "source_sha256": "23f9e9264d22a027c651251162e79dae000f8084b0682cc4a24e06267297f7c3",
      "locator": "L64–93, base price table",
      "evidence_excerpt": "Sonnet 5.5",
      "scope": "Claude Sonnet 5.5 first-party standard API rates",
      "conditions": "5-minute cache write $2.50, 1-hour $4, cache hit $0.10 per MTok. Cache writes and hits have separate rates. Different model from the GPT-OSS cost illustration.",
      "input_rate": 2,
      "output_rate": 10,
      "model": "claude-sonnet-5.5",
      "mode": "firstparty",
      "claim_sha256": "d541f0a6ecdf6c15f1f0889f7255e7f7d19b5994ddb1e4d6a0fbcbc3aeee649e"
    },
    {
      "id": "IA-020",
      "vendor": "anthropic",
      "field": "batch",
      "label": "Message Batches",
      "value": "50% off input and output",
      "source_id": "anthropic-pricing",
      "source_url": "https://platform.claude.com/docs/en/about-claude/pricing",
      "source_role": "vendor_claim",
      "read_at": "2026-11-14T13:15:04.957Z",
      "source_sha256": "23f9e9264d22a027c651251162e79dae000f8084b0682cc4a24e06267297f7c3",
      "locator": "L219–254, Batch processing",
      "evidence_excerpt": "50 percent discount",
      "scope": "Claude Message Batches",
      "conditions": "Batch and Fast cannot be combined. Partner regional pricing and model-specific long-context conditions can differ.",
      "claim_sha256": "3c16086e06bb691bdc9c84be2181f82425297d4dc98792d91d1d0b04de9152d3"
    },
    {
      "id": "IA-021",
      "vendor": "anthropic",
      "field": "data",
      "label": "Fable / Mythos 5.2",
      "value": "30-day retention required unless expressly authorized",
      "source_id": "anthropic-release-notes",
      "source_url": "https://docs.anthropic.com/en/release-notes/api",
      "source_role": "vendor_claim",
      "read_at": "2026-11-13T04:25:04.098Z",
      "source_sha256": "3ab83d9b5a2f50cdd1ebf8e7747bcc9519083ada9a7814c3e033fc376ebc7271",
      "locator": "text lines 13–21, November 9, 2026",
      "evidence_excerpt": "Both models require 30-day data retention",
      "scope": "Claude Fable 5.2 and Claude Mythos 5.2",
      "conditions": "Not available under ZDR unless Anthropic expressly authorizes it. Do not generalize to every Claude model.",
      "claim_sha256": "8a321c99730598c2248fe879776385f86048e874f81deeb370bf1bc76ca0c392"
    },
    {
      "id": "IA-022",
      "vendor": "anthropic",
      "field": "features",
      "label": "Fable 5.2 migration",
      "value": "Changing an earlier prefix can cause a 400 error",
      "source_id": "anthropic-release-notes",
      "source_url": "https://docs.anthropic.com/en/release-notes/api",
      "source_role": "vendor_claim",
      "read_at": "2026-11-13T04:25:04.098Z",
      "source_sha256": "3ab83d9b5a2f50cdd1ebf8e7747bcc9519083ada9a7814c3e033fc376ebc7271",
      "locator": "text lines 14–20, November 9, 2026",
      "evidence_excerpt": "the request returns a 400 error",
      "scope": "Fable/Mythos 5.2 thinking blocks reused after system/tools/earlier-message edits",
      "conditions": "Every account subject to prefix check. Earlier computer_20251124 tool also returns 400 except on Bedrock; migrate to computer_toolset_20260801. Model switch does not guarantee API feature parity.",
      "claim_sha256": "965166c40ab192b1e851a6ba9b4e1cf452bc2b0dbf676b5c9c4041c8fa35fb16"
    },
    {
      "id": "IA-023",
      "vendor": "google-vertex-ai",
      "field": "price",
      "label": "Flash introductory tokens",
      "value": "$0.75 input / $3.75 output per 1M tokens",
      "source_id": "google-genai-pricing",
      "source_url": "https://cloud.google.com/vertex-ai/generative-ai/pricing",
      "source_role": "vendor_claim",
      "read_at": "2026-11-13T04:24:59.939Z",
      "source_sha256": "ac25524147e3d5dacb7109dcc12a6448085aa9ca218a354d8af4017e0234a2ce",
      "locator": "text lines 1–5, introductory pricing",
      "evidence_excerpt": "through December 31, 2026",
      "scope": "Gemini 3.8, 3.7 and 3.6 Flash; CodeMender using these models",
      "conditions": "USD introductory rates through December 31, 2026. Announced January 1, 2027 standard rates $1.50 input / $7.50 output. Separate model from the GPT-OSS illustration.",
      "input_rate": 0.75,
      "output_rate": 3.75,
      "model": "Gemini Flash 3.8/3.7/3.6",
      "mode": "cloud",
      "claim_sha256": "e58641cf4ed8076ba4f105694bb0039c1f20d1d0315885b798384fd523912313"
    },
    {
      "id": "IA-024",
      "vendor": "google-vertex-ai",
      "field": "billing",
      "label": "Failed long completions",
      "value": "Gemini 4 retrievable usage can still be charged",
      "source_id": "google-genai-pricing",
      "source_url": "https://cloud.google.com/vertex-ai/generative-ai/pricing",
      "source_role": "vendor_claim",
      "read_at": "2026-11-13T04:24:59.939Z",
      "source_sha256": "ac25524147e3d5dacb7109dcc12a6448085aa9ca218a354d8af4017e0234a2ce",
      "locator": "text lines 88–91, Gemini 4 footnote",
      "evidence_excerpt": "When a completion exceeds 32K tokens",
      "scope": "Gemini 4 completion above 32K tokens",
      "conditions": "Effective November 13, 2026: retrievable token usage is charged even if the request ultimately fails due to client/server timeout. This is an exception to the page’s general non-200 statement.",
      "claim_sha256": "5742941c5077d757823179ccd4155b609060bf46bbd07f8e117f77247c343fe6"
    },
    {
      "id": "IA-025",
      "vendor": "google-vertex-ai",
      "field": "data",
      "label": "Grounding retention",
      "value": "Search up to 3 days; Maps 30 days",
      "source_id": "google-data",
      "source_url": "https://docs.cloud.google.com/gemini-enterprise-agent-platform/resources/zero-data-retention",
      "source_role": "vendor_claim",
      "read_at": "2026-11-14T13:22:23.484Z",
      "source_sha256": "2caec635f1414a071abbcdcc2554cb08cc9143be166baef4457bd29b8ced7ff3",
      "locator": "L59–68",
      "evidence_excerpt": "There is no way to disable",
      "scope": "Grounding with Google Search and Grounding with Google Maps",
      "conditions": "Search logs include derived queries/context; Maps stores prompts/context/output. Cannot disable this storage while using those features. Advanced AI features may have additional logging; no blanket ZDR claim.",
      "claim_sha256": "9e7c56272e7569d5137032cf1ce7dfafe2bd4a00541c20d09c80f8ad86806f5c"
    },
    {
      "id": "IA-026",
      "vendor": "google-vertex-ai",
      "field": "data",
      "label": "Interactions API",
      "value": "store defaults to true",
      "source_id": "google-data",
      "source_url": "https://docs.cloud.google.com/gemini-enterprise-agent-platform/resources/zero-data-retention",
      "source_role": "vendor_claim",
      "read_at": "2026-11-14T13:22:23.484Z",
      "source_sha256": "2caec635f1414a071abbcdcc2554cb08cc9143be166baef4457bd29b8ced7ff3",
      "locator": "L67–70",
      "evidence_excerpt": "it defaults to `true` for all models",
      "scope": "Managed models through Interactions API",
      "conditions": "Explicit store=false needed for that application state. Request-response BigQuery logging is separately off by default; other retention features still need review.",
      "claim_sha256": "fb1a0a7cf35166abbba13eba5b39061b16b889c5897a12cbadcf6373a875b307"
    },
    {
      "id": "IA-027",
      "vendor": "amazon-bedrock",
      "field": "capacity",
      "label": "Account quotas",
      "value": "Per-model and per-region; account quota may be lower",
      "source_id": "bedrock-quotas",
      "source_url": "https://docs.aws.amazon.com/bedrock/latest/userguide/quotas.html",
      "source_role": "vendor_claim",
      "read_at": "2026-11-13T04:25:14.281Z",
      "source_sha256": "248bacbf3fb484bbdb1390a39978581ab55fadb83080d85180b204686da9209f",
      "locator": "Quotas paragraph and model list",
      "evidence_excerpt": "quota",
      "scope": "Amazon Bedrock Runtime / Mantle",
      "conditions": "Embedding models may be RPM-only. Account age, payment and usage history can affect quotas; no automatic increase or common throughput promise.",
      "claim_sha256": "d9a38b0743588debdc5e83c7375a013795f5f9a0271759430244feeeb4b0f43b"
    },
    {
      "id": "IA-028",
      "vendor": "amazon-bedrock",
      "field": "data",
      "label": "Retention mode",
      "value": "none blocks a model that requires retention",
      "source_id": "bedrock-retention",
      "source_url": "https://docs.aws.amazon.com/bedrock/latest/userguide/data-retention.html",
      "source_role": "vendor_claim",
      "read_at": "2026-11-14 13:32:35 UTC",
      "source_sha256": "22238527132e0c21800de0c03a0584f32389174d67d3732a1028c095f67d0fb6",
      "locator": "L8–27, L145–149",
      "evidence_excerpt": "will block the request and return an error",
      "scope": "Account per Region; project setting only on bedrock-mantle",
      "conditions": "New scope inherits model default. store=false alone is not ZDR. none prohibits durable request/response storage; Responses store=true rejected and background mode unavailable.",
      "claim_sha256": "3a2acc8ba3da003be6fd12e70321a3c740b5cfe8152a720fc4bcd5bbaeac1366"
    },
    {
      "id": "IA-029",
      "vendor": "amazon-bedrock",
      "field": "data",
      "label": "Fable 5 / 5.1 / 5.2 on Bedrock",
      "value": "AWS may retain and review content for up to 30 days",
      "source_id": "bedrock-retention-details",
      "source_url": "https://docs.aws.amazon.com/bedrock/latest/userguide/data-retention.html",
      "source_role": "vendor_claim",
      "read_at": "2026-11-14 13:32:44 UTC",
      "source_sha256": "b2a2a077e315a2a030aec9a64f404c024ce5615ae4b6bad3ffa6910239ab555f",
      "locator": "L256–263",
      "evidence_excerpt": "retained within the AWS boundary for up to 30 days",
      "scope": "Models requiring aws_review, currently Claude Fable 5, 5.1, 5.2",
      "conditions": "Retained within the AWS boundary; not shared with model provider. Cross-region inference stores retained inputs/outputs in destination processing Regions. ZDR exceptions require per-account/per-model eligibility.",
      "also_fields": [
        "region"
      ],
      "claim_sha256": "a2ee0e98047662072b234d950adf19082f95a9725040dbd9e49e168ef94ad17b"
    },
    {
      "id": "IA-030",
      "vendor": "microsoft-azure-ai-foundry",
      "field": "region",
      "label": "Processing scope",
      "value": "Global / Data Zone / Azure geography",
      "source_id": "microsoft-deployments",
      "source_url": "https://learn.microsoft.com/en-us/azure/foundry/foundry-models/concepts/deployment-types",
      "source_role": "vendor_claim",
      "read_at": "2026-11-14T13:14:31.421Z",
      "source_sha256": "2fc68d284e76bed2041d4560509699a8d95e68940fcb626c61f8c314d86c52ce",
      "locator": "L31–47",
      "evidence_excerpt": "Global types: May be processed in any Azure region",
      "scope": "Foundry Serverless API deployment types, not managed compute",
      "conditions": "At-rest data stays in designated Azure geography. Global can process any Azure region; Data Zone within US/EU/APAC zone; Standard/Regional Provisioned within geography, possibly between its regions.",
      "mode": "cloud",
      "claim_sha256": "072cf12845c1fe31b8d0e4461f97db4d54f6010f46bde0f7b3ba85c018f129cb"
    },
    {
      "id": "IA-031",
      "vendor": "microsoft-azure-ai-foundry",
      "field": "batch",
      "label": "Global Batch",
      "value": "50% discount; 24-hour processing",
      "source_id": "microsoft-deployments",
      "source_url": "https://learn.microsoft.com/en-us/azure/foundry/foundry-models/concepts/deployment-types",
      "source_role": "vendor_claim",
      "read_at": "2026-11-14T13:14:31.421Z",
      "source_sha256": "2fc68d284e76bed2041d4560509699a8d95e68940fcb626c61f8c314d86c52ce",
      "locator": "L57–72, deployment table",
      "evidence_excerpt": "50% discount, 24-hr",
      "scope": "Global Batch, selected supported models",
      "conditions": "Processing may occur in any Azure region. Data Zone Batch stays within its zone. Not every model supports every deployment type.",
      "claim_sha256": "e18c39946cbce2f039de7c70f0844f5c1bcdeeb12f464899b13e4b6986c69c28"
    },
    {
      "id": "IA-032",
      "vendor": "microsoft-azure-ai-foundry",
      "field": "capacity",
      "label": "Reserved capacity",
      "value": "Provisioned types bill reserved PTU",
      "source_id": "microsoft-deployments",
      "source_url": "https://learn.microsoft.com/en-us/azure/foundry/foundry-models/concepts/deployment-types",
      "source_role": "vendor_claim",
      "read_at": "2026-11-14T13:14:31.421Z",
      "source_sha256": "2fc68d284e76bed2041d4560509699a8d95e68940fcb626c61f8c314d86c52ce",
      "locator": "L57–72",
      "evidence_excerpt": "Reserved PTU",
      "scope": "Global / Data Zone / Regional Provisioned",
      "conditions": "Model availability and capacity differ by type. Developer evaluation is a separate 24-hour-lifetime type with no SLA or data residency guarantee.",
      "claim_sha256": "0601e0a890d76bc5054cac5807aa2acfff8834dae4c6c94ec15941fdb9dccb0e"
    },
    {
      "id": "IA-033",
      "vendor": "openrouter",
      "field": "price",
      "label": "Platform fee",
      "value": "Standard 5.5% / Business 8%",
      "source_id": "openrouter-pricing",
      "source_url": "https://openrouter.ai/pricing",
      "source_role": "vendor_claim",
      "read_at": "2026-11-14T13:14:18.585Z",
      "source_sha256": "299fa3e03c128ee9718d29b5a68d30051055aa62d8ba60beeb06e20f4a38ab59",
      "locator": "L50–84, plan table",
      "evidence_excerpt": "Standard 5.5%",
      "scope": "OpenRouter paid plans, underlying model token rates separate",
      "conditions": "Model tokens billed at posted per-model input/output rates. BYOK has separate fee thresholds. This is not a per-token rate or application total.",
      "mode": "gateway",
      "claim_sha256": "cdf046eb7580db8939ed78722e201a67db7515a0119fadfc0904418cf507a258"
    },
    {
      "id": "IA-034",
      "vendor": "openrouter",
      "field": "region",
      "label": "In-region routing",
      "value": "Business and Enterprise; US/EU",
      "source_id": "openrouter-pricing",
      "source_url": "https://openrouter.ai/pricing",
      "source_role": "vendor_claim",
      "read_at": "2026-11-14T13:14:18.585Z",
      "source_sha256": "299fa3e03c128ee9718d29b5a68d30051055aa62d8ba60beeb06e20f4a38ab59",
      "locator": "L64–76",
      "evidence_excerpt": "Standard No  | Business Yes",
      "scope": "OpenRouter US/EU in-region routing",
      "conditions": "Standard has data-policy routing but not these in-region features. Contractual SLA Enterprise only; paid rate limits pass through underlying provider.",
      "also_fields": [
        "capacity"
      ],
      "claim_sha256": "40431cc9b53303bc9fe374a8104505fed93804050cfe27d59c9798d7047149aa"
    },
    {
      "id": "IA-035",
      "vendor": "openrouter",
      "field": "data",
      "label": "Content retention",
      "value": "Opt-in and safety/legal exceptions",
      "source_id": "openrouter-data",
      "source_url": "https://openrouter.ai/docs/guides/privacy/data-collection",
      "source_role": "vendor_claim",
      "read_at": "2026-11-14T13:22:23.484Z",
      "source_sha256": "285063dc584ab2e5992b9ffa2c2f56cd0b2055095a2151837e6806ade3e5912b",
      "locator": "L151–163, prompt logging",
      "evidence_excerpt": "Safety Retention",
      "scope": "OpenRouter prompt/response logging",
      "conditions": "Logging/product improvement off by default. Safety Retention endpoints can retain any request for abuse monitoring; exclude them with ZDR privacy settings. CSAM/legal exceptions and metadata remain.",
      "claim_sha256": "14ecaa1a8b82059152ddd9087622906ec119c075cc1a362d2105469e3955bf9f"
    },
    {
      "id": "IA-036",
      "vendor": "replicate",
      "field": "billing",
      "label": "Public vs private serving",
      "value": "Private models can bill setup and idle time",
      "source_id": "replicate-pricing",
      "source_url": "https://replicate.com/pricing",
      "source_role": "vendor_claim",
      "read_at": "2026-11-14T13:14:18.585Z",
      "source_sha256": "898a553c68fe1d818679b74f81804a022845c9f8f4b0780f6bc26478794fc167",
      "locator": "L8–25, public/private billing",
      "evidence_excerpt": "the time they spend idle, waiting for requests",
      "scope": "Replicate private models except fast-booting fine-tunes",
      "conditions": "Public models bill processing by time/hardware or input/output. Private models bill setup+idle+active while online; fast-booting fine-tunes are an explicit active-only exception.",
      "mode": "host",
      "claim_sha256": "d03ea17eda44ca9b9544610aa0e894e4ec643ceb2ba06cf028fa0710d48c1531"
    },
    {
      "id": "IA-037",
      "vendor": "cerebras",
      "field": "capacity",
      "label": "GPT-OSS-120B PayGo limits",
      "value": "500K uncached TPM / 1K RPM",
      "source_id": "cerebras-paygo",
      "source_url": "https://support.cerebras.net/articles/5041581099-cerebras-self-serve-paygo-faq",
      "source_role": "vendor_claim",
      "read_at": "2026-11-14 13:33:47 UTC",
      "source_sha256": "e8ebe6c1fc7a876e5acb5996815e81641bc56fa4df2fd1585d752513eb107717",
      "locator": "L34–47",
      "evidence_excerpt": "Uncached Tokens Per Minute (TPM): 500,000",
      "scope": "Cerebras Developer PayGo, GPT-OSS-120B",
      "conditions": "Models and limits may change without notice. Custom higher limits, SLAs and dedicated capacity require agreement; no production suitability inferred from a self-serve limit.",
      "mode": "host",
      "claim_sha256": "4268450c8c686e85f306201efe93a34c167fe2125305c1a50a1e2954052cd048"
    },
    {
      "id": "IA-038",
      "vendor": "cerebras",
      "field": "billing",
      "label": "Prepaid credits",
      "value": "Purchased credits expire after one year",
      "source_id": "cerebras-paygo",
      "source_url": "https://support.cerebras.net/articles/5041581099-cerebras-self-serve-paygo-faq",
      "source_role": "vendor_claim",
      "read_at": "2026-11-14 13:33:47 UTC",
      "source_sha256": "e8ebe6c1fc7a876e5acb5996815e81641bc56fa4df2fd1585d752513eb107717",
      "locator": "L49–62",
      "evidence_excerpt": "credits expire one year after purchase",
      "scope": "Cerebras PayGo purchased credits",
      "conditions": "At zero balance rate limits revert to zero until funds added. The separate promotional-credit terms are in IA-040.",
      "claim_sha256": "a3f0e94b09c8ecbc7a7baf1826759e1b3d95aa02c297eb80f742a6fef0283a31"
    },
    {
      "id": "IA-039",
      "vendor": "cerebras",
      "field": "features",
      "label": "Completion choices",
      "value": "Only n=1 supported",
      "source_id": "cerebras-chat",
      "source_url": "https://inference-docs.cerebras.ai/api-reference/chat-completions",
      "source_role": "vendor_claim",
      "read_at": "2026-11-14 13:33:47 UTC",
      "source_sha256": "2c3bfa85b201a370893126718dd80890523163bbc4f5f60f5bab2fe02df61ccc",
      "locator": "L344–354",
      "evidence_excerpt": "Cerebras currently supports only `1`",
      "scope": "Cerebras Chat Completions",
      "conditions": "max_completion_tokens includes reasoning. Reasoning effort defaults and support vary by model; compatible message format does not imply full API parity.",
      "claim_sha256": "47c968599ad6f1ba4094d8809b5dbb8dd51730e2f51bdb74cbb3f08027c9046d"
    },
    {
      "id": "IA-040",
      "vendor": "cerebras",
      "field": "billing",
      "label": "Promotional credit",
      "value": "$5 once; expires after 30 days",
      "source_id": "cerebras-pricing",
      "source_url": "https://www.cerebras.ai/pricing",
      "source_role": "vendor_claim",
      "read_at": "2026-11-14T13:14:09.865Z",
      "source_sha256": "81d64c2890d319aefe78010e3e10be0bc175d4c6599e351fc7b1e6d0e8349306",
      "locator": "L22–25, pricing footnotes",
      "evidence_excerpt": "one-time $5 promotional credit",
      "scope": "Cerebras promotional trial with valid payment method",
      "conditions": "No automatic paid enrollment. API/Playground access pauses when exhausted or expired unless PayGo credits separately purchased. Numeric tier table was not exposed in this opened text; do not treat this as a current production quote.",
      "claim_sha256": "f836fabc8ccdfa2a54ec5973e7cd4630605940e3dc71c6b6eae9684212665a03"
    },
    {
      "id": "IA-041",
      "vendor": "openai",
      "field": "capacity",
      "label": "GPT-6.1 Sol usage tiers",
      "value": "Build: 5,000 RPM / 1,000,000 TPM",
      "source_id": "openai-sol",
      "source_url": "https://developers.openai.com/api/docs/models/gpt-6.1-sol",
      "source_role": "vendor_claim",
      "read_at": "2026-11-14 13:32:44 UTC",
      "source_sha256": "39f22f9fbca0e062165e8a6c896c0d2e3a37f125b6c21abd244ec131be100016",
      "locator": "L1197–1205, Tier / RPM / TPM table",
      "evidence_excerpt": "Build  | 5,000  | 1,000,000",
      "scope": "gpt-6.1-sol API usage-tier limits",
      "conditions": "Launch lists 10,000 RPM / 4,000,000 TPM; Grow 15,000 RPM / 40,000,000 TPM. Free unsupported. Usage/spend can raise the tier; actual account limits and achieved throughput were not checked.",
      "claim_sha256": "57f2a8636923461d4d3f6eda8ddfdc99d2332d65c8f6ceaf95efa8faee3c5b6f"
    }
  ],
  "sources": [
    {
      "id": "openai-api-data-use",
      "vendor": "openai",
      "url": "https://developers.openai.com/api/docs/guides/your-data",
      "read_at": "2026-11-13T04:25:05.526Z",
      "role": "vendor_claim",
      "snapshot": "02-source/snapshots/openai-api-data-use-17d804a75634e3dbeae598f81998deb0ac471a61710ed5b64bd594bb6d855580.txt",
      "sha256": "17d804a75634e3dbeae598f81998deb0ac471a61710ed5b64bd594bb6d855580",
      "capture_scope": "Previously saved extracted text reused, not refreshed"
    },
    {
      "id": "together-privacy",
      "vendor": "together-ai",
      "url": "https://docs.together.ai/docs/privacy-and-security",
      "read_at": "2026-11-13T04:25:08.899Z",
      "role": "vendor_claim",
      "snapshot": "02-source/snapshots/together-privacy-0b709b5615d0112c857094e19eb78cd84c71cc32a80913476a07dd9c1ddac50b.txt",
      "sha256": "0b709b5615d0112c857094e19eb78cd84c71cc32a80913476a07dd9c1ddac50b",
      "capture_scope": "Previously saved extracted text reused, not refreshed"
    },
    {
      "id": "together-rate-limits",
      "vendor": "together-ai",
      "url": "https://docs.together.ai/docs/serverless/rate-limits",
      "read_at": "2026-11-13T04:25:07.967Z",
      "role": "vendor_claim",
      "snapshot": "02-source/snapshots/together-rate-limits-0e468d21026e02e9d4715b6dceda4d45497bb3578cbdc0edc14a68bb1c806b7d.txt",
      "sha256": "0e468d21026e02e9d4715b6dceda4d45497bb3578cbdc0edc14a68bb1c806b7d",
      "capture_scope": "Previously saved extracted text reused, not refreshed"
    },
    {
      "id": "bedrock-quotas",
      "vendor": "amazon-bedrock",
      "url": "https://docs.aws.amazon.com/bedrock/latest/userguide/quotas.html",
      "read_at": "2026-11-13T04:25:14.281Z",
      "role": "vendor_claim",
      "snapshot": "02-source/snapshots/bedrock-quotas-248bacbf3fb484bbdb1390a39978581ab55fadb83080d85180b204686da9209f.txt",
      "sha256": "248bacbf3fb484bbdb1390a39978581ab55fadb83080d85180b204686da9209f",
      "capture_scope": "Previously saved extracted text reused, not refreshed"
    },
    {
      "id": "google-genai-pricing",
      "vendor": "google-vertex-ai",
      "url": "https://cloud.google.com/vertex-ai/generative-ai/pricing",
      "read_at": "2026-11-13T04:24:59.939Z",
      "role": "vendor_claim",
      "snapshot": "02-source/snapshots/google-genai-pricing-ac25524147e3d5dacb7109dcc12a6448085aa9ca218a354d8af4017e0234a2ce.txt",
      "sha256": "ac25524147e3d5dacb7109dcc12a6448085aa9ca218a354d8af4017e0234a2ce",
      "capture_scope": "Previously saved extracted text reused, not refreshed"
    },
    {
      "id": "anthropic-release-notes",
      "vendor": "anthropic",
      "url": "https://docs.anthropic.com/en/release-notes/api",
      "read_at": "2026-11-13T04:25:04.098Z",
      "role": "vendor_claim",
      "snapshot": "02-source/snapshots/anthropic-release-notes-3ab83d9b5a2f50cdd1ebf8e7747bcc9519083ada9a7814c3e033fc376ebc7271.txt",
      "sha256": "3ab83d9b5a2f50cdd1ebf8e7747bcc9519083ada9a7814c3e033fc376ebc7271",
      "capture_scope": "Previously saved extracted text reused, not refreshed"
    },
    {
      "id": "fireworks-pricing",
      "vendor": "fireworks-ai",
      "url": "https://fireworks.ai/pricing",
      "read_at": "2026-11-14T13:14:09.865Z",
      "role": "vendor_claim",
      "snapshot": "02-source/fresh/snapshots/fireworks-pricing.txt",
      "sha256": "1d3a58dcf7e8c99c65881527493424c7c65bd0876cbe0bd20f8a738df43f80e2",
      "raw": "02-source/fresh/raw/open-1.json",
      "raw_sha256": "e73008bb2d135b3fa1fbc936fe390c682ad1753fa3abfbe96b863be5539c4a5a",
      "capture_scope": "Full web tool response retained; selected rendered passages, not full HTTP archive"
    },
    {
      "id": "cerebras-pricing",
      "vendor": "cerebras",
      "url": "https://www.cerebras.ai/pricing",
      "read_at": "2026-11-14T13:14:09.865Z",
      "role": "vendor_claim",
      "snapshot": "02-source/fresh/snapshots/cerebras-pricing.txt",
      "sha256": "81d64c2890d319aefe78010e3e10be0bc175d4c6599e351fc7b1e6d0e8349306",
      "raw": "02-source/fresh/raw/open-1.json",
      "raw_sha256": "e73008bb2d135b3fa1fbc936fe390c682ad1753fa3abfbe96b863be5539c4a5a",
      "capture_scope": "Full web tool response retained; selected rendered passages, not full HTTP archive"
    },
    {
      "id": "groq-models",
      "vendor": "groq",
      "url": "https://console.groq.com/docs/models",
      "read_at": "2026-11-14T13:14:18.585Z",
      "role": "vendor_claim",
      "snapshot": "02-source/fresh/snapshots/groq-models.txt",
      "sha256": "29e04ae4326f939560e3a9d0d8acbdda0d3d43ab95f88916f1266702338c5821",
      "raw": "02-source/fresh/raw/open-2.json",
      "raw_sha256": "83b802942772e90abcd57286d3a1a75929a9fec1c2ce137eb9044bd43d6eeba8",
      "capture_scope": "Full web tool response retained; selected rendered passages, not full HTTP archive"
    },
    {
      "id": "openrouter-pricing",
      "vendor": "openrouter",
      "url": "https://openrouter.ai/pricing",
      "read_at": "2026-11-14T13:14:18.585Z",
      "role": "vendor_claim",
      "snapshot": "02-source/fresh/snapshots/openrouter-pricing.txt",
      "sha256": "299fa3e03c128ee9718d29b5a68d30051055aa62d8ba60beeb06e20f4a38ab59",
      "raw": "02-source/fresh/raw/open-2.json",
      "raw_sha256": "83b802942772e90abcd57286d3a1a75929a9fec1c2ce137eb9044bd43d6eeba8",
      "capture_scope": "Full web tool response retained; selected rendered passages, not full HTTP archive"
    },
    {
      "id": "replicate-pricing",
      "vendor": "replicate",
      "url": "https://replicate.com/pricing",
      "read_at": "2026-11-14T13:14:18.585Z",
      "role": "vendor_claim",
      "snapshot": "02-source/fresh/snapshots/replicate-pricing.txt",
      "sha256": "898a553c68fe1d818679b74f81804a022845c9f8f4b0780f6bc26478794fc167",
      "raw": "02-source/fresh/raw/open-2.json",
      "raw_sha256": "83b802942772e90abcd57286d3a1a75929a9fec1c2ce137eb9044bd43d6eeba8",
      "capture_scope": "Full web tool response retained; selected rendered passages, not full HTTP archive"
    },
    {
      "id": "anthropic-pricing",
      "vendor": "anthropic",
      "url": "https://platform.claude.com/docs/en/about-claude/pricing",
      "read_at": "2026-11-14T13:15:04.957Z",
      "role": "vendor_claim",
      "snapshot": "02-source/fresh/snapshots/anthropic-pricing.txt",
      "sha256": "23f9e9264d22a027c651251162e79dae000f8084b0682cc4a24e06267297f7c3",
      "raw": "02-source/fresh/raw/find-1.json",
      "raw_sha256": "2e4ccb7d0af0b5968120c81c70eed46e01a166cc1a64efddf1662a9960a6a06d",
      "capture_scope": "Full web tool response retained; selected rendered passages, not full HTTP archive"
    },
    {
      "id": "microsoft-deployments",
      "vendor": "microsoft-azure-ai-foundry",
      "url": "https://learn.microsoft.com/en-us/azure/foundry/foundry-models/concepts/deployment-types",
      "read_at": "2026-11-14T13:14:31.421Z",
      "role": "vendor_claim",
      "snapshot": "02-source/fresh/snapshots/microsoft-deployments.txt",
      "sha256": "2fc68d284e76bed2041d4560509699a8d95e68940fcb626c61f8c314d86c52ce",
      "raw": "02-source/fresh/raw/open-3.json",
      "raw_sha256": "17d250f2e0e1ef2bb158488c3fd29b4437b11097e7863bd694e9f10413372b0c",
      "capture_scope": "Full web tool response retained; selected rendered passages, not full HTTP archive"
    },
    {
      "id": "deepinfra-gptoss",
      "vendor": "deepinfra",
      "url": "https://deepinfra.com/openai/gpt-oss-120b",
      "read_at": "2026-11-14T13:14:45.928Z",
      "role": "vendor_claim",
      "snapshot": "02-source/fresh/snapshots/deepinfra-gptoss.txt",
      "sha256": "c92a8429151c33d33085919a3522f70acf24ece488e81162c62a69182adb12d4",
      "raw": "02-source/fresh/raw/open-4.json",
      "raw_sha256": "c5e8f881f262f4dcbc978712a4dfa02ea664bcca6068d194b677efa80c5fd672",
      "capture_scope": "Full web tool response retained; selected rendered passages, not full HTTP archive"
    },
    {
      "id": "google-data",
      "vendor": "google-vertex-ai",
      "url": "https://docs.cloud.google.com/gemini-enterprise-agent-platform/resources/zero-data-retention",
      "read_at": "2026-11-14T13:22:23.484Z",
      "role": "vendor_claim",
      "snapshot": "02-source/fresh/snapshots/google-data.txt",
      "sha256": "2caec635f1414a071abbcdcc2554cb08cc9143be166baef4457bd29b8ced7ff3",
      "raw": "02-source/fresh/raw/open-5.json",
      "raw_sha256": "ca3446926b661645ee3dab4b91c4203e57561ddcf2e58c55fd0d57e8c2571b24",
      "capture_scope": "Full web tool response retained; selected rendered passages, not full HTTP archive"
    },
    {
      "id": "groq-data",
      "vendor": "groq",
      "url": "https://console.groq.com/docs/your-data",
      "read_at": "2026-11-14T13:22:23.484Z",
      "role": "vendor_claim",
      "snapshot": "02-source/fresh/snapshots/groq-data.txt",
      "sha256": "a6e86c4777cd1e703a70cb2fe6937872afeff3a540fff524bc1f013f6a27942d",
      "raw": "02-source/fresh/raw/open-5.json",
      "raw_sha256": "ca3446926b661645ee3dab4b91c4203e57561ddcf2e58c55fd0d57e8c2571b24",
      "capture_scope": "Full web tool response retained; selected rendered passages, not full HTTP archive"
    },
    {
      "id": "openrouter-data",
      "vendor": "openrouter",
      "url": "https://openrouter.ai/docs/guides/privacy/data-collection",
      "read_at": "2026-11-14T13:22:23.484Z",
      "role": "vendor_claim",
      "snapshot": "02-source/fresh/snapshots/openrouter-data.txt",
      "sha256": "285063dc584ab2e5992b9ffa2c2f56cd0b2055095a2151837e6806ade3e5912b",
      "raw": "02-source/fresh/raw/open-5.json",
      "raw_sha256": "ca3446926b661645ee3dab4b91c4203e57561ddcf2e58c55fd0d57e8c2571b24",
      "capture_scope": "Full web tool response retained; selected rendered passages, not full HTTP archive"
    },
    {
      "id": "bedrock-retention",
      "vendor": "amazon-bedrock",
      "url": "https://docs.aws.amazon.com/bedrock/latest/userguide/data-retention.html",
      "read_at": "2026-11-14 13:32:35 UTC",
      "role": "vendor_claim",
      "snapshot": "02-source/fresh/snapshots/bedrock-retention.txt",
      "sha256": "22238527132e0c21800de0c03a0584f32389174d67d3732a1028c095f67d0fb6",
      "raw": "02-source/fresh/raw/open-7.json",
      "raw_sha256": "5cb7499907cb6d91b62bdf3ba7b343543a5cfe64d80028c88333cf16a384ad1f",
      "capture_scope": "Full web tool response retained; selected rendered passages, not full HTTP archive"
    },
    {
      "id": "bedrock-retention-details",
      "vendor": "amazon-bedrock",
      "url": "https://docs.aws.amazon.com/bedrock/latest/userguide/data-retention.html",
      "read_at": "2026-11-14 13:32:44 UTC",
      "role": "vendor_claim",
      "snapshot": "02-source/fresh/snapshots/bedrock-retention-details.txt",
      "sha256": "b2a2a077e315a2a030aec9a64f404c024ce5615ae4b6bad3ffa6910239ab555f",
      "raw": "02-source/fresh/raw/open-8.json",
      "raw_sha256": "8657716792bc8ddc814ea492356882f3190066f929bb175348f717ad3734851d",
      "capture_scope": "Full web tool response retained; selected rendered passages, not full HTTP archive"
    },
    {
      "id": "openai-sol",
      "vendor": "openai",
      "url": "https://developers.openai.com/api/docs/models/gpt-6.1-sol",
      "read_at": "2026-11-14 13:32:44 UTC",
      "role": "vendor_claim",
      "snapshot": "02-source/fresh/snapshots/openai-sol.txt",
      "sha256": "39f22f9fbca0e062165e8a6c896c0d2e3a37f125b6c21abd244ec131be100016",
      "raw": "02-source/fresh/raw/open-8.json",
      "raw_sha256": "8657716792bc8ddc814ea492356882f3190066f929bb175348f717ad3734851d",
      "capture_scope": "Full web tool response retained; selected rendered passages, not full HTTP archive"
    },
    {
      "id": "together-gptoss",
      "vendor": "together-ai",
      "url": "https://www.together.ai/models/gpt-oss-120b",
      "read_at": "2026-11-14 13:32:44 UTC",
      "role": "vendor_claim",
      "snapshot": "02-source/fresh/snapshots/together-gptoss.txt",
      "sha256": "3120d139ec354ff83df343a2ff4b9f360d64169dae8b8350e954e8138af81eba",
      "raw": "02-source/fresh/raw/open-8.json",
      "raw_sha256": "8657716792bc8ddc814ea492356882f3190066f929bb175348f717ad3734851d",
      "capture_scope": "Full web tool response retained; selected rendered passages, not full HTTP archive"
    },
    {
      "id": "cerebras-paygo",
      "vendor": "cerebras",
      "url": "https://support.cerebras.net/articles/5041581099-cerebras-self-serve-paygo-faq",
      "read_at": "2026-11-14 13:33:47 UTC",
      "role": "vendor_claim",
      "snapshot": "02-source/fresh/snapshots/cerebras-paygo.txt",
      "sha256": "e8ebe6c1fc7a876e5acb5996815e81641bc56fa4df2fd1585d752513eb107717",
      "raw": "02-source/fresh/raw/open-9.json",
      "raw_sha256": "da59b92fb7c9a852081633e3e65a6567e4532ce4fcc7a6f23486e6083495e65d",
      "capture_scope": "Full web tool response retained; selected rendered passages, not full HTTP archive"
    },
    {
      "id": "cerebras-chat",
      "vendor": "cerebras",
      "url": "https://inference-docs.cerebras.ai/api-reference/chat-completions",
      "read_at": "2026-11-14 13:33:47 UTC",
      "role": "vendor_claim",
      "snapshot": "02-source/fresh/snapshots/cerebras-chat.txt",
      "sha256": "2c3bfa85b201a370893126718dd80890523163bbc4f5f60f5bab2fe02df61ccc",
      "raw": "02-source/fresh/raw/open-9.json",
      "raw_sha256": "da59b92fb7c9a852081633e3e65a6567e4532ce4fcc7a6f23486e6083495e65d",
      "capture_scope": "Full web tool response retained; selected rendered passages, not full HTTP archive"
    }
  ],
  "prior_facts": "03-aggregate/facts.csv",
  "views": {
    "industry_baseline": [
      "price",
      "billing",
      "capacity",
      "batch",
      "data",
      "region",
      "features"
    ],
    "feedback_led": [
      "billing",
      "capacity",
      "data",
      "region",
      "features"
    ]
  },
  "history": [
    {
      "id": "H-IA-01",
      "kind": "announced_change",
      "effective_at": "2027-01-01",
      "text": "Google announces Flash introductory-rate end; future rates are not today’s rate.",
      "claims": [
        "IA-023"
      ]
    },
    {
      "id": "H-IA-02",
      "kind": "dated_release",
      "effective_at": "2026-11-09",
      "text": "Fable 5.2 introduces prefix/tool migration conditions and retains model-specific 30-day retention.",
      "claims": [
        "IA-021",
        "IA-022"
      ]
    },
    {
      "id": "H-IA-03",
      "kind": "dated_condition",
      "effective_at": "2026-11-13",
      "text": "Google documents retrievable usage charges for failed Gemini 4 completions above 32K tokens.",
      "claims": [
        "IA-024"
      ]
    }
  ],
  "unknowns": [
    "Independent output fidelity, serving precision for every host, latency distribution and reliability",
    "Live account quotas, all model/region availability and negotiated terms",
    "DeepInfra full endpoint retention policy, not just its model badge",
    "Cerebras current numeric token table not exposed in opened pricing text; production tier mapping not independently established",
    "Actual multi-step application tokens, retries, tool fees and cache-hit rates",
    "No recurring refresh scheduler or measured commercial outcome"
  ],
  "recommendations": [
    {
      "id": "REC-IA-01",
      "audience": "Production buyer moving an OpenAI-shaped client",
      "action": "Check model and feature limits before switching the endpoint.",
      "supporting_facts": [
        "IA-016",
        "IA-022",
        "IA-039"
      ],
      "limitations": "Documented differences; not a compatibility test."
    },
    {
      "id": "REC-IA-02",
      "audience": "Buyer requiring request-content controls",
      "action": "Check the selected model and optional feature, then the retention switch.",
      "supporting_facts": [
        "IA-012",
        "IA-025",
        "IA-028",
        "IA-035"
      ],
      "limitations": "Vendor statements, not a security certification or legal advice."
    }
  ]
}
