{
  "id": 4089182,
  "title": "We said a Claude Code subagent costs 436k tokens. A cleaner measurement says 54k — here is what fooled us",
  "url": "https://urgent.news/2026/08/29/we-said-a-claude-code-subagent-costs-436k-tokens-a-cleaner",
  "topic": "ai",
  "section": "AI",
  "published": "2026-08-29T02:17:00.000Z",
  "source": {
    "name": "Dev.to",
    "slug": "dev-to",
    "url": "https://dev.to/rulestack/we-said-a-claude-code-subagent-costs-436k-tokens-a-cleaner-measurement-says-54k-here-is-what-37am"
  },
  "original_language": "en",
  "account": "Two weeks ago, a number was published stating that a Claude Code subagent costs around 436,000 tokens before it reads a single file. The same measurement method was repeated in two articles, but today a cleaner measurement yielded a different result: 54,154 tokens. This discrepancy is more significant than the actual number, as it exposes the flaw in the initial approach. The original method involved running the same review task with multiple agents and attributing the difference to per-agent overhead. However, this method was flawed as it counted the same 54,154 context tokens multiple times, inflating the actual spawning cost. The cleaner measurement involved spawning a minimal subagent with a simple prompt, resulting in a single request containing 54,154 tokens in the agent's transcript file. This revealed that the original 436,000 token cost was not the actual spawning cost but rather the cost multiplied by the number of times the context was thought to be paid for. The break-even analysis was redone, showing that delegating a read becomes cheaper than doing it in the main loop when the inline reading cost exceeds approximately 33,000 tokens. This correction has important implications for measuring the cost of using Claude Code subagents, as it highlights the importance of accurate measurement methods to avoid overestimating the true cost.",
  "summary": "Two weeks ago we published a number: a Claude Code subagent costs ~436,000 tokens before it reads a single file. We built routing rules around it, wrote a break-even formula with it, and repeated it in two articles. Today we re-measured with a cleaner method and got 54,154 tokens . The old number was off by roughly 8×, and the way it was wrong is more useful than the number itself. This post is…",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 4,
    "also_reported_by": [
      {
        "outlet": "XDA Developers",
        "title": "Claude Code was using 51,000 tokens before I even typed a prompt — I fixed it",
        "url": "https://urgent.news/2026/08/28/claude-code-was-using-51-000-tokens-before-i-even-typed-a-prompt-i",
        "published": "2026-08-28T20:00:15.000Z"
      },
      {
        "outlet": "The Register Science",
        "title": "Researcher shows how Claude Code can be tricked simply by asking it to summarize a website",
        "url": "https://urgent.news/2026/08/28/researcher-shows-how-claude-code-can-be-tricked-simply-by-asking-it",
        "published": "2026-08-28T20:50:15.000Z"
      },
      {
        "outlet": "Dev.to",
        "title": "Using Claude Code or Codex as a Teacher, Not an Autopilot",
        "url": "https://urgent.news/2026/08/29/using-claude-code-or-codex-as-a-teacher-not-an-autopilot",
        "published": "2026-08-29T04:54:34.000Z"
      }
    ]
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}