{
  "schema": "varexis-public-evidence-record/0.3",
  "generated_at_utc": "2026-10-28T23:55:00.000Z",
  "record_id": "VXI-EXT-ICE04-OPENAI-AGENT-TOKEN-SUBSETS-01-v1",
  "record_kind": "external_reference",
  "title": "OpenAI Agents token subsets and double counting",
  "system": "OpenAI Agents API",
  "evidence_family": "vendor_evidence_review",
  "index_status": "PUBLIC_SAFE_SOURCE_REVIEW",
  "evidence_state": "PRIMARY_DOCUMENTATION_REVIEW_ONLY",
  "canonical_question": "Should reasoning and cached tokens be added to OpenAI Agents input and output totals?",
  "near_questions": [
    "Are reasoning tokens already included in Agents output_tokens?",
    "Are cached tokens additional to input_tokens?",
    "How can a token-cost spreadsheet avoid counting detail fields twice?",
    "Do token subtotals alone establish an agent invoice?"
  ],
  "supported_claims": [
    {
      "claim": "OpenAI documents cached tokens as a subset of input_tokens and reasoning tokens as a subset of output_tokens in Agents usage. Adding those detailed categories again to the parent counts double-counts recorded tokens. Preserve the subsets for applicable pricing analysis while keeping the totals and component categories distinct.",
      "reliance": "DOCUMENTED_CAPABILITY_ONLY"
    }
  ],
  "not_demonstrated_claims": [
    "Token arithmetic alone does not establish the final charge, complete task cost or applicable model rates.",
    "Buyer implementation, live runtime enforcement, settled payment, certification or permission to act are not established.",
    "No expanded capabilities or licences for existing INDEX executables."
  ],
  "public_summary": "OpenAI documents cached tokens as a subset of input_tokens and reasoning tokens as a subset of output_tokens in Agents usage. Adding those detailed categories again to the parent counts double-counts recorded tokens. Preserve the subsets for applicable pricing analysis while keeping the totals and component categories distinct.",
  "related_record_ids": [],
  "source_review": {
    "source_type": "official_external_documentation",
    "source_title": "OpenAI Agents API: Observability and usage",
    "source_url": "https://developers.openai.com/api/docs/guides/agents-api/observability.md",
    "checked_on": "2026-10-28",
    "candidate_record_id": "VXI-EXT-ICE04-OPENAI-AGENT-TOKEN-SUBSETS-01-v1",
    "candidate_record_sha256": "b66bf82f658bcc42382e0ce50cda87a1c54bafabf741825dba84259bb4c32503",
    "method": "Primary documentation reviewed for the published bounded statement.",
    "deployment_tested_by_varexis": false,
    "revalidation": "The delivered snapshot preserves the review date shown. Recheck current primary documentation and deployment context before relying on present applicability; payment does not trigger fresh research."
  },
  "answer_policy": {
    "classification": "DOCUMENTED_WITH_DEPLOYMENT_BOUNDARY",
    "bounded_answer": "OpenAI documents cached tokens as a subset of input_tokens and reasoning tokens as a subset of output_tokens in Agents usage. Adding those detailed categories again to the parent counts double-counts recorded tokens. Preserve the subsets for applicable pricing analysis while keeping the totals and component categories distinct.",
    "required_buyer_context": "Usage object with parent and detail counts; Model and accounting scope; Applicable pricing and billing evidence. Buyer deployment not inspected.",
    "missing_evidence": "A broader claim needs evidence matching the intended use. Payment does not upgrade the evidence state."
  },
  "private_boundary": "Raw logs, thresholds, private paths, source code, protected mechanisms and reconstruction-level details are excluded from public and paid delivery.",
  "decision_profile": {
    "schema": "varexis-vendor-evidence-review/1",
    "review_status": "documentation_review_only",
    "checked_on": "2026-10-28",
    "claim_to_check": "Should reasoning and cached tokens be added to OpenAI Agents input and output totals?",
    "documented_statement": {
      "status": "documented",
      "statement": "OpenAI documents cached tokens as a subset of input_tokens and reasoning tokens as a subset of output_tokens in Agents usage. Adding those detailed categories again to the parent counts double-counts recorded tokens. Preserve the subsets for applicable pricing analysis while keeping the totals and component categories distinct.",
      "source_ids": [
        "openai-agent-usage"
      ]
    },
    "review_inference": {
      "status": "review_inference",
      "statement": "Review openai agents token subsets and double counting before accepting the buyer claim. Parent/subset arithmetic, including reasoning output, is distinct from deciding cached versus uncached rates. This is a documentation-review judgment, not an observed deployment outcome.",
      "source_ids": [
        "openai-agent-usage"
      ]
    },
    "buyer_evidence_required": [
      "Usage object with parent and detail counts",
      "Model and accounting scope",
      "Applicable pricing and billing evidence"
    ],
    "review_condition": {
      "status": "varexis_review_policy",
      "statement": "Token arithmetic alone does not establish the final charge, complete task cost or applicable model rates."
    },
    "unknowns": "Buyer configuration, effective authority, complete usage and final business outcomes were not inspected.",
    "unknown_finding": {
      "status": "unknown",
      "statement": "Buyer configuration, effective authority, complete usage and final business outcomes were not inspected."
    },
    "evidence_sources": [
      {
        "source_id": "openai-agent-usage",
        "url": "https://developers.openai.com/api/docs/guides/agents-api/observability.md",
        "final_url": "https://developers.openai.com/api/docs/guides/agents-api/observability.md",
        "status": 200,
        "captured_at_utc": "2026-10-28T23:49:42.485491+00:00",
        "sha256": "92e5dbfe555fbfaa2119ee30f24b84aaba65197237f18b99d89ba8bd23276754",
        "bytes": 18786,
        "content_type": "text/markdown; charset=utf-8",
        "title": "OpenAI Agents API: Observability and usage",
        "hash_basis": "SHA-256 of captured HTTP response body bytes",
        "checked_on": "2026-10-28",
        "capture_status": 200,
        "claim_review_completed": true,
        "reviewed_sections": [
          "Understand token usage"
        ],
        "version_context": "Rolling official documentation captured 2026-10-28; exact response hash is pinned, not a stable vendor release. Verify the deployed API/version and account feature availability. Agents API beta documentation; this is not a qualification of the buyer implementation.",
        "rights_review": "Original factual paraphrase and attributed source metadata only. Captured documentation and code remain private; no source prose, examples, images or vendor marks are redistributed. No licence to vendor software or buyer assurance is granted."
      }
    ],
    "observed_deployment": false,
    "commercial_reliance_boundary": "Source documentation supports only the stated boundary. Existing licensed checkers analyse supported supplied declarations; they do not query these providers, verify live systems, settle payments or certify compliance.",
    "paid_value": {
      "decision_context": "Review openai agents token subsets and double counting before accepting the buyer claim. Parent/subset arithmetic, including reasoning output, is distinct from deciding cached versus uncached rates.",
      "free": "Complete review, practical questions, source references and limitations; freely accessible and locally saveable.",
      "single_request": "Unavailable for this publication-only record; historical saved retrieval remains unchanged.",
      "evidence_receipt": "Unavailable for this publication-only record; no new paid retrieval entitlement.",
      "excluded": [
        "Fresh vendor research",
        "Live provider inspection",
        "Runtime enforcement or certification",
        "Expanded executable schemas or licences",
        "Executable AI cost reconciliation"
      ],
      "integrity": "Hashes identify content versions; they do not prove truth or provide a signature."
    },
    "revalidation": {
      "trigger": "Recheck the source and applicable API, account and configuration before current reliance. No monitoring or fresh research is included.",
      "specific_watch": "Understand token usage"
    },
    "buyer_relevance": "Review openai agents token subsets and double counting before accepting the buyer claim. Parent/subset arithmetic, including reasoning output, is distinct from deciding cached versus uncached rates.",
    "refusal_conditions": [
      "Token arithmetic alone does not establish the final charge, complete task cost or applicable model rates.",
      "Do not infer buyer deployment conformance or successful business outcomes from source documentation."
    ]
  },
  "niche": {
    "niche_id": "ice04-ai-cost-usage-invoice-assurance",
    "label": "AI Cost, Usage & Invoice Assurance",
    "classification_id": "ai-cost-usage-invoice-assurance",
    "classification_label": "AI Cost, Usage & Invoice Assurance",
    "topic_id": "ai-cost-usage-invoice-assurance",
    "buyer_intent": "Review openai agents token subsets and double counting before accepting the buyer claim. Parent/subset arithmetic, including reasoning output, is distinct from deciding cached versus uncached rates.",
    "comparison_role": "supplementary_discovery"
  },
  "paid_access_note": "The full published answer and boundaries are free. Paid access saves the selected route and source versions, with an unsigned receipt when purchased. Preview coverage before paying. Fresh research, bespoke assessment and protected evidence are excluded.",
  "service_route": {
    "request_guide": "/varexis-index/request.html",
    "request_workflow": "/varexis-index/request-workflow.json",
    "service_id": "single_request",
    "service_routes_url": "/varexis-index/service-routes.json",
    "payment_required": true,
    "availability_endpoint": "/api/index/capabilities",
    "operational_status": "AUTOMATIC_SERVICE_CAPABILITIES_GATED",
    "checkout_url": "/index-access"
  },
  "authority_boundary": "VAREXIS-INDEX informs evidence reliance. It does not provide advice, certification, deployment authority, security assurance, production approval, or guarantee.",
  "human_page": "/varexis-index/catalog.html#openai-agent-token-subsets",
  "machine_readable_record": "/varexis-index/records/openai-agent-token-subsets.json"
}