PulseAugur
EN
LIVE 20:21:12

New tool reveals Anthropic Claude Code subagents re-send 51K tokens per call

By PulseAugur Editorial · [1 sources] ·

A developer has created a Python tool called `subagent-tax` to analyze the token usage of Anthropic's Claude Code subagents. The tool reveals that each subagent call re-sends a fixed preamble of approximately 51,000 tokens, which includes the system prompt, tool schemas, and skill listings. This can lead to significant token waste, with an estimated $20.20 in costs for 132 observed calls in one user's case. The tool helps identify the most costly components of this preamble, suggesting optimizations like slimming down the system prompt to save tokens. AI

IMPACT Highlights potential inefficiencies in LLM agent architectures, prompting developers to optimize token usage and reduce costs.

RANK_REASON The cluster describes a new, third-party tool developed to analyze the performance and cost-efficiency of an existing AI product's features.

Read on dev.to — Claude Code tag →

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →

New tool reveals Anthropic Claude Code subagents re-send 51K tokens per call

How we ranked this

Signal score
5 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Tool
The cluster describes a new, third-party tool developed to analyze the performance and cost-efficiency of an existing AI product's features.
Source corroboration
Single-source cluster
Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics
product, infra
Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance
High
Clearly on-topic for AI-industry coverage.
Story freshness
Same-day
Cluster formed today. Ranking reflects the current source set at time of score.

Full methodology in our editorial standards.

COVERAGE [1]

  1. dev.to — Claude Code tag TIER_1 English(EN) · hao li ·

    Your Subagents Are Re-Sending 51K Tokens Every Call. I Built a Tool to Prove It.

    <p>A recent post here by <a class="mentioned-user" href="https://dev.to/ji_ai">@ji_ai</a> measured something startling: Claude Code subagents were <strong>48% of their bill</strong> while producing <strong>0.9% of the output</strong>. The culprit: every <code>Task()</code> call r…

RELATED ENTITIES

RELATED TOPICS