Count · Measurement specification

Measurement plan

Which token-counting requests were made?

What this proves

The plan defines every count_tokens call and verifies that the projected full request round-trips to the captured body.

Call set: 3 overview + 1 prompt + 40 item = 44
Projection check: Passed

Input: Countable captured request
Output: Named count_tokens call set
File: provenance/count-tokens-plan.json

Inspect the measurement specification

JSON

View raw

{
  "generatedAt": "2026-06-09T06:05:57.379Z",
  "endpoint": "https://api.anthropic.com/v1/messages/count_tokens?beta=true",
  "model": "claude-opus-4-8",
  "capturedAnthropicBeta": "claude-code-20250219,context-1m-2025-08-07,interleaved-thinking-2025-05-14,context-management-2025-06-27,prompt-caching-scope-2026-01-05,mid-conversation-system-2026-04-07,advisor-tool-2026-03-01,effort-2025-11-24",
  "countTokensBeta": "token-counting-2024-11-01",
  "countableFields": [
    "model",
    "messages",
    "thinking",
    "output_config",
    "system",
    "tools"
  ],
  "deliberatelyExcludedNonContextFields": [
    "context_management",
    "max_tokens",
    "metadata",
    "stream"
  ],
  "overviewCategories": [
    "built_in_tool_definitions",
    "skill_descriptions",
    "everything_else"
  ],
  "categoryCalls": 3,
  "skillIndependentEstimateCalls": 13,
  "toolIndependentEstimateCalls": 27,
  "promptAblationCalls": 1,
  "totalPlannedCalls": 44,
  "fullBodyCanonicalSha256": "d87aa5f349109324f0cc80267cdc6d8a5750cb414e50293972791e834bd3d8e4",
  "fullBodyRoundTripsCapturedProjection": true
}