# Claude Code Token Usage Statistics (2026)

> How many tokens Claude Code actually burns in 2026: per-session input/output volumes, the input-dominated cost split, cache economics, and inter-model spread — every figure sourced from primary documentation and published research.

Source: https://tokenade.net/en/stats/claude-code-token-usage-statistics · Updated: 2026-10-16

Claude Code's token bill comes down to a handful of numbers: how many tokens a session actually consumes, the lopsided split between what it reads and what it writes, the discount caching gives you on repeated context, and the per-million-token rate of the model behind it. This page collects those numbers from primary sources so you don't have to reconstruct them from a billing dashboard at 2 a.m.

I build token-reduction tooling for a living, so I read these analyses with a vested interest — and the headline I keep coming back to is unglamorous: the expensive thing is what the agent *reads*, not what it writes. Prices, cache multipliers and the default model were re-checked against Anthropic's documentation as of October 2026; model versions and rates move fast, so re-verify at the primary source before you quote them in a contract.

## Key Takeaways

<div className="my-8 p-6 bg-amber-50 border-l-4 border-amber-500 rounded-r-lg">
  <h3 className="text-lg font-bold text-amber-800 mb-4">Claude Code token usage at a glance</h3>
  <ul className="space-y-3">
    <li className="flex items-start gap-2">
      <span className="text-amber-600 font-bold">•</span>
      <span>A typical 50-turn agentic session burns roughly <strong>1,000,000 input tokens and 40,000 output tokens</strong> — a 25:1 ratio. <a href="#source-1" className="text-blue-600 hover:underline">[1]</a></span>
    </li>
    <li className="flex items-start gap-2">
      <span className="text-amber-600 font-bold">•</span>
      <span>Input tokens drive <strong>≈85% of total session cost</strong> — the agent re-reads its whole context on every tool call. <a href="#source-1" className="text-blue-600 hover:underline">[1]</a></span>
    </li>
    <li className="flex items-start gap-2">
      <span className="text-amber-600 font-bold">•</span>
      <span>Uber burned through its <strong>entire 2026 AI budget in four months</strong> after Claude Code adoption reached 84% across 5,000 engineers, at <strong>$500–$2,000 per heavy user per month</strong>. <a href="#source-2" className="text-blue-600 hover:underline">[2]</a></span>
    </li>
    <li className="flex items-start gap-2">
      <span className="text-amber-600 font-bold">•</span>
      <span>Prompt-cache reads cost <strong>5% of fresh input on Opus 5.5</strong> (Claude Code's default) and 10% on Sonnet 5.5 and Haiku — a 90–95% saving on repeated context. <a href="#source-5" className="text-blue-600 hover:underline">[5]</a></span>
    </li>
    <li className="flex items-start gap-2">
      <span className="text-amber-600 font-bold">•</span>
      <span>Re-sent context is <strong>≈62% of the bill</strong> at production teams audited in 2026. <a href="#source-4" className="text-blue-600 hover:underline">[4]</a></span>
    </li>
    <li className="flex items-start gap-2">
      <span className="text-amber-600 font-bold">•</span>
      <span>On <strong>Opus 5.5 ($4/$20 per MTok)</strong>, Claude Code's default, that 50-turn session costs around <strong>$4.80</strong> with no caching; on Sonnet 5.5 ($2/$10), around <strong>$2.40</strong>. <a href="#source-5" className="text-blue-600 hover:underline">[5]</a></span>
    </li>
    <li className="flex items-start gap-2">
      <span className="text-amber-600 font-bold">•</span>
      <span>Anthropic's own enterprise average is <strong>≈$13 per developer per active day</strong> ($150–$250 a month), with 90% of users under $30 a day. <a href="#source-11" className="text-blue-600 hover:underline">[11]</a></span>
    </li>
  </ul>
</div>

> Lower your own Claude Code numbers: Tokenade compacts command output and file reads before they reach the model.
>
> Install: `npm install -g @tokenade/cli && tokenade install` — or let your agent do it: https://tokenade.net/install.md

## How many tokens does a Claude Code session consume?

A typical 50-turn agentic coding session consumes roughly **1,000,000 input tokens and 40,000 output tokens** — a 25:1 input-to-output ratio. <a href="#source-1" className="text-blue-600 hover:underline">[1]</a> That ratio is the single most important fact about Claude Code's economics, and it surprises almost everyone the first time they see it.

The mechanism is straightforward once you say it out loud. Claude Code is an agent: every tool call (read a file, run a test, grep the repo) ships the *entire accumulated context* back to the model — system prompt, tool definitions, conversation history, file contents, command output. By turn 30 of a debugging session, the agent is re-reading 25,000–35,000 tokens of history on every single request. The output it writes — a patch, a paragraph of explanation — is tiny by comparison. <a href="#source-1" className="text-blue-600 hover:underline">[1]</a>

What does that add up to per person? Anthropic's own Claude Code documentation puts the average across enterprise deployments at **around $13 per developer per active day and $150–$250 per developer per month**, with costs staying **below $30 per active day for 90% of users**. <a href="#source-11" className="text-blue-600 hover:underline">[11]</a> The long tail is where the stories come from: running several instances in parallel or wiring the agent into automation moves you out of that 90% fast. If your monthly bill looks like a salary line item, the [context window](https://tokenade.net/en/glossary/context-window) accumulation pattern above is why.

<div className="my-6 grid grid-cols-1 md:grid-cols-2 gap-4">
  <div className="p-6 bg-gradient-to-br from-blue-50 to-blue-100 rounded-lg border border-blue-200 text-center">
    <div className="text-4xl font-bold text-blue-700">25:1</div>
    <div className="text-sm text-blue-600 mt-2">input-to-output token ratio in a typical 50-turn session</div>
    <div className="text-xs text-gray-500 mt-2">Source: Vantage, 2026 <a href="#source-1" className="text-blue-600 hover:underline">[1]</a></div>
  </div>
  <div className="p-6 bg-gradient-to-br from-orange-50 to-orange-100 rounded-lg border border-orange-200 text-center">
    <div className="text-4xl font-bold text-orange-700">≈85%</div>
    <div className="text-sm text-orange-600 mt-2">of total session cost driven by input, not output</div>
    <div className="text-xs text-gray-500 mt-2">Source: Vantage, 2026 <a href="#source-1" className="text-blue-600 hover:underline">[1]</a></div>
  </div>
</div>

## Why is input the expensive side, when output costs 5× more per token?

This is the part that trips people up. On every Claude tier, output tokens are priced at **5× the input rate** — $50 vs $10 per million on Fable 5.1, $20 vs $4 on Opus 5.5, $10 vs $2 on Sonnet 5.5, $5 vs $1 on Haiku 4.5. <a href="#source-5" className="text-blue-600 hover:underline">[5]</a> So the naive instinct is to make the agent write less.

It's the wrong lever. Input tokens drive **approximately 85% of total session cost** in production, because the volume gap dwarfs the price gap. <a href="#source-1" className="text-blue-600 hover:underline">[1]</a> A 25:1 volume ratio against a 5:1 price ratio means input still wins the bill by a factor of five. Telling Claude Code to "be concise" trims the cheapest, smallest part of your spend.

The high-leverage move is the opposite: control what the agent *reads* on each turn. Retrieve only the relevant code instead of dumping whole files, prune stale conversation history, and don't re-feed the model output it already saw. That is precisely the gap [output filtering](https://tokenade.net/en/glossary/output-filtering) and [semantic code search](https://tokenade.net/en/glossary/semantic-code-search) are built to close.

## How much do the tokens actually cost?

The per-million-token rate depends on which model is driving Claude Code. Since the Opus 5.5 / Sonnet 5.5 / Haiku 5.5 generation, the current Claude lineup spans a 100× range on input price, from Fable 5.1 at the top to Haiku 5.5 at the bottom. Claude Code's `default` model setting resolves to Opus 5.5 on Pro, Max, Team, Enterprise and the Anthropic API. <a href="#source-12" className="text-blue-600 hover:underline">[12]</a>

<div className="my-8 overflow-x-auto">
  <table className="min-w-full border-collapse border border-gray-300">
    <thead>
      <tr className="bg-gray-100">
        <th className="border border-gray-300 px-4 py-3 text-left font-semibold">Model</th>
        <th className="border border-gray-300 px-4 py-3 text-center font-semibold">Input ($/MTok)</th>
        <th className="border border-gray-300 px-4 py-3 text-center font-semibold">Output ($/MTok)</th>
        <th className="border border-gray-300 px-4 py-3 text-center font-semibold">Cache read</th>
        <th className="border border-gray-300 px-4 py-3 text-center font-semibold">≈ cost of one 50-turn session</th>
      </tr>
    </thead>
    <tbody>
      <tr>
        <td className="border border-gray-300 px-4 py-2 font-medium">Claude Fable 5.1</td>
        <td className="border border-gray-300 px-4 py-2 text-center">$10.00</td>
        <td className="border border-gray-300 px-4 py-2 text-center">$50.00</td>
        <td className="border border-gray-300 px-4 py-2 text-center">2.5% of input</td>
        <td className="border border-gray-300 px-4 py-2 text-center">≈$12.00</td>
      </tr>
      <tr className="bg-gray-50">
        <td className="border border-gray-300 px-4 py-2 font-medium">Claude Opus 5.5 (Claude Code default)</td>
        <td className="border border-gray-300 px-4 py-2 text-center">$4.00</td>
        <td className="border border-gray-300 px-4 py-2 text-center">$20.00</td>
        <td className="border border-gray-300 px-4 py-2 text-center">5% of input</td>
        <td className="border border-gray-300 px-4 py-2 text-center">≈$4.80</td>
      </tr>
      <tr>
        <td className="border border-gray-300 px-4 py-2 font-medium">Claude Sonnet 5.5</td>
        <td className="border border-gray-300 px-4 py-2 text-center">$2.00</td>
        <td className="border border-gray-300 px-4 py-2 text-center">$10.00</td>
        <td className="border border-gray-300 px-4 py-2 text-center">10% of input</td>
        <td className="border border-gray-300 px-4 py-2 text-center">≈$2.40</td>
      </tr>
      <tr className="bg-gray-50">
        <td className="border border-gray-300 px-4 py-2 font-medium">Claude Haiku 4.5</td>
        <td className="border border-gray-300 px-4 py-2 text-center">$1.00</td>
        <td className="border border-gray-300 px-4 py-2 text-center">$5.00</td>
        <td className="border border-gray-300 px-4 py-2 text-center">10% of input</td>
        <td className="border border-gray-300 px-4 py-2 text-center">≈$1.20</td>
      </tr>
      <tr>
        <td className="border border-gray-300 px-4 py-2 font-medium">Claude Haiku 5.5 (prompts up to 100K tokens)</td>
        <td className="border border-gray-300 px-4 py-2 text-center">$0.10</td>
        <td className="border border-gray-300 px-4 py-2 text-center">$0.50</td>
        <td className="border border-gray-300 px-4 py-2 text-center">10% of input</td>
        <td className="border border-gray-300 px-4 py-2 text-center">≈$0.12</td>
      </tr>
    </tbody>
  </table>
  <p className="text-xs text-gray-500 mt-2">Session cost assumes ≈1M input + 40K output tokens, no caching. Haiku 5.5 bills $0.50 / $2.50 per MTok on prompts over 100K tokens. Prices: Anthropic, Claude API Pricing docs (platform.claude.com, checked October 2026) <a href="#source-5" className="text-blue-600 hover:underline">[5]</a>; session volume from Vantage <a href="#source-1" className="text-blue-600 hover:underline">[1]</a>.</p>
</div>

For the full cross-vendor picture — OpenAI's and Google's lineups included — see the [LLM API token pricing](https://tokenade.net/en/stats/llm-api-token-pricing) reference and the [AI coding agent token costs](https://tokenade.net/en/stats/ai-coding-agent-token-costs) breakdown. The detail that matters here: cache reads on Anthropic are no longer a flat 10%. They are billed at **10% of the base input price on Sonnet and Haiku, 5% on Opus 5.5 and 2.5% on Fable 5.1**. <a href="#source-5" className="text-blue-600 hover:underline">[5]</a>

## How much does prompt caching change the Claude Code bill?

Anthropic [prompt caching](https://tokenade.net/en/glossary/prompt-caching) bills cache-read tokens at **10% of the base input price** on most models — **5% on Opus 5.5, 2.5% on Fable 5.1** — while a cache write costs **1.25× the base input price** for a 5-minute TTL, or **2× for a 1-hour TTL**. The default lifetime is 5 minutes, refreshed each time the prefix is reused. <a href="#source-3" className="text-blue-600 hover:underline">[3]</a> <a href="#source-5" className="text-blue-600 hover:underline">[5]</a>

The catch is that the cached prefix must be byte-identical across requests. Reorder one instruction and the cache misses. GitHub measured the prefix problem in its own agentic workflows in May 2026: the most common waste was unused MCP tool registrations (a 40-tool GitHub MCP server adds 10–15 KB of schema to every turn), and after its fixes, cost-weighted "effective tokens" fell between 19% and 62% depending on the workflow, while one workflow rose 5%. <a href="#source-6" className="text-blue-600 hover:underline">[6]</a> When the prefix drifts — which is the default for many agent harnesses — you pay full freight.

Caching helps with the *stable* prefix. It does nothing for the part of the bill that is genuinely new-but-redundant: the same file re-read three turns apart, the test output you already acted on, the directory listing you've now seen five times. That residue is the bulk of what [context compression](https://tokenade.net/en/glossary/context-compression) targets.

## How much of the Claude Code bill is just re-sent context?

A lot. A 2026 LeanOps audit of 30 engineering teams running agentic AI in production found that **re-sent context accounts for ≈62% of the total bill**. <a href="#source-4" className="text-blue-600 hover:underline">[4]</a> The same audit reported that teams hitting 50–70% cost reductions within two weeks consistently combined per-user budget caps, prompt caching, model-tier routing, and context-window pruning. <a href="#source-4" className="text-blue-600 hover:underline">[4]</a>

That is the headline number for anyone trying to cut a Claude Code bill: roughly six of every ten dollars is the model re-reading things it has already seen. You cannot fix that by switching models or writing terser prompts — only by changing what gets sent.

<div className="my-8 p-6 bg-white rounded-lg border border-gray-200 shadow-sm">
  <p className="text-lg font-semibold mb-6 text-center">Where the Claude Code bill actually goes</p>
  <div className="space-y-4 overflow-hidden">
    <div>
      <div className="flex justify-between mb-1">
        <span className="text-sm font-medium">Re-sent / redundant context</span>
        <span className="text-sm font-semibold">≈62% of bill</span>
      </div>
      <div className="w-full bg-gray-200 rounded-full h-4 overflow-hidden">
        <div className="bg-blue-600 h-4 rounded-full" style={{width: '62%', maxWidth: '100%'}}></div>
      </div>
    </div>
    <div>
      <div className="flex justify-between mb-1">
        <span className="text-sm font-medium">Net-new input (relevant code, fresh tool output)</span>
        <span className="text-sm font-semibold">remainder of input</span>
      </div>
      <div className="w-full bg-gray-200 rounded-full h-4 overflow-hidden">
        <div className="bg-green-600 h-4 rounded-full" style={{width: '23%', maxWidth: '100%'}}></div>
      </div>
    </div>
    <div>
      <div className="flex justify-between mb-1">
        <span className="text-sm font-medium">Output (patches, explanations)</span>
        <span className="text-sm font-semibold">≈15% of cost</span>
      </div>
      <div className="w-full bg-gray-200 rounded-full h-4 overflow-hidden">
        <div className="bg-gray-400 h-4 rounded-full" style={{width: '15%', maxWidth: '100%'}}></div>
      </div>
    </div>
  </div>
  <p className="text-xs text-gray-500 mt-4">Sources: LeanOps audit <a href="#source-4" className="text-blue-600 hover:underline">[4]</a>, Vantage analysis <a href="#source-1" className="text-blue-600 hover:underline">[1]</a> (2026). Bar widths are qualitative rankings, not a single computed metric.</p>
</div>

## What does Claude Code actually cost at scale?

Enterprise and individual data points from 2026 give a concrete sense of how token spend scales with usage intensity.

**Uber** burned through its entire 2026 AI tools budget by April — four months in — after Claude Code adoption jumped from 32% to 84% across its 5,000-engineer workforce. The average engineer cost **$150–$250 a month**; heavy users were running up **$500–$2,000 per engineer per month**, and Uber's leadership acknowledged they could not yet quantify the productivity return despite around 70% of committed code coming from AI tools. <a href="#source-2" className="text-blue-600 hover:underline">[2]</a>

**Peter Steinberger**'s team ran roughly 100 Codex instances in the cloud for a month, working on the OpenClaw project, and posted a bill of **$1.3 million — 603 billion tokens across 7.6 million requests**. One caveat worth carrying: the exact figures come from a screenshot in Steinberger's own post rather than a published statement. But the numbers remain the most concrete public benchmark for autonomous agentic coding at scale. <a href="#source-7" className="text-blue-600 hover:underline">[7]</a>

**A Reddit developer** left a Claude Code loop running overnight — polling for software updates every 30 minutes — and woke up to a **$6,000 bill**. The culprit was a combination of an 800,000-token context window being rebuilt from scratch on every cycle (Claude Code had started writing many requests with a 5-minute cache TTL instead of 1 hour in early March 2026; Anthropic later said in the GitHub thread that the client now picks the TTL per request to lower total cost) and no spending cap. There was no real-time dashboard to warn him; the first alert was the email confirming the damage. <a href="#source-8" className="text-blue-600 hover:underline">[8]</a>

At the other end of the scale, Henry Godnick documented a **surprise $80 bill** from a batch-processing script he left running unattended — a reminder that even small loops compound fast when each iteration rebuilds a multi-turn context. <a href="#source-9" className="text-blue-600 hover:underline">[9]</a>

These data points share a common structure: the cost spike is not a single large model call but many medium calls, each re-sending a growing context. The practical implication is that the [savings dashboard](https://tokenade.net/en) matters most precisely when a session is long or a loop is running — the exact moments when you're least likely to be watching.

## What does research say about agent context management costs?

A study by Lindenbauer et al. (JetBrains Research), arXiv:2508.21433, first posted August 2025 and presented at the DL4C workshop at NeurIPS 2025, benchmarked context-management strategies for LLM-based software engineering agents on SWE-bench Verified across five model configurations. <a href="#source-10" className="text-blue-600 hover:underline">[10]</a>

The headline finding is counter-intuitive: **simple observation masking** (truncating older tool outputs) halves cost relative to the raw agent baseline while *matching* the solve rate of more expensive LLM-based summarisation. With Qwen3-Coder 480B, observation masking cut cost per instance by 52.7% ($1.29 to $0.61) against the raw baseline while the solve rate went from 53.4% to 54.8%. The October 2025 revision adds a hybrid of masking and summarisation that trims cost by a further 7% versus masking alone. The study challenges the assumption that smarter context management necessarily means more LLM calls — sometimes it just means not sending old data.

The broader implication for Claude Code users: the agent harness design (what it keeps, what it discards, how it handles long tool outputs) is as important a cost variable as the model tier you choose.

## Methodology note

Per-token prices are from Anthropic's official pricing page (primary source, verified October 2026). Prompt-caching multipliers are from Anthropic's pricing and prompt caching documentation (primary source). The per-developer daily and monthly averages and the default model come from Anthropic's Claude Code documentation (primary source, October 2026). Session-volume and cost-split figures are from the Vantage production analysis. The 62% re-sent-context figure is from an industry audit of 30 production teams, not an academic study, and should be treated as indicative. Enterprise cost data (Uber, OpenClaw) is from contemporaneous press reporting. The Lindenbauer et al. arXiv preprint (2508.21433) is the academic source for context-management cost benchmarks on SWE-bench Verified. Indicative session costs use the published Anthropic rates with the Vantage volume profile and no caching applied. Re-verify every figure at its primary source before citing in commercial contexts.

<div className="my-8 p-6 bg-slate-50 rounded-lg border border-slate-200">
  <h3 className="text-lg font-bold mb-2">Want these numbers to go down?</h3>
  <p className="text-sm text-slate-700 mb-3">Tokenade sits in front of Claude Code (and Cursor, Codex, Copilot, Windsurf) and cuts the input side automatically — semantic code search instead of whole-file dumps, output filtering, skeleton compression, and MCP result compaction — and `tokenade dashboard` opens a graphical report of exactly what you saved. Free and open source (MIT/Apache-2.0); no cap. Compaction runs entirely on your own machine.</p>
  <p className="text-sm"><a href="/en" className="text-blue-600 font-medium hover:underline">See how Tokenade reduces Claude Code token usage →</a></p>
</div>

<div className="my-8 p-6 bg-gray-50 rounded-lg" id="sources">
  <h2 className="text-xl font-bold mb-4">Sources and references</h2>
  <ol className="space-y-3 text-sm">
    <li className="flex gap-2" id="source-1">
      <span className="font-semibold text-gray-500">[1]</span>
      <span>Vantage. "The Hidden Cost Driver in Agentic Coding Sessions in 2026". vantage.sh, 2026. <a href="https://www.vantage.sh/blog/agentic-coding-costs" className="text-blue-600 hover:underline ml-1" target="_blank" rel="noopener noreferrer">Link ↗</a></span>
    </li>
    <li className="flex gap-2" id="source-2">
      <span className="font-semibold text-gray-500">[2]</span>
      <span>Storyboard18 / The Information. "Uber exhausts 2026 AI budget in four months amid massive Claude Code adoption". May 2026. <a href="https://www.storyboard18.com/brand-marketing/uber-exhausts-2026-ai-budget-in-four-months-amid-massive-claude-code-adoption-98443.htm" className="text-blue-600 hover:underline ml-1" target="_blank" rel="noopener noreferrer">Link ↗</a></span>
    </li>
    <li className="flex gap-2" id="source-3">
      <span className="font-semibold text-gray-500">[3]</span>
      <span>Anthropic. "Prompt caching — Claude API Docs". platform.claude.com, 2026. <a href="https://platform.claude.com/docs/en/build-with-claude/prompt-caching" className="text-blue-600 hover:underline ml-1" target="_blank" rel="noopener noreferrer">Link ↗</a></span>
    </li>
    <li className="flex gap-2" id="source-4">
      <span className="font-semibold text-gray-500">[4]</span>
      <span>LeanOps. "AI Agents Burn 50x More Tokens Than Chats". leanopstech.com, 2026. <a href="https://leanopstech.com/blog/agentic-ai-cost-runaway-token-budget-2026/" className="text-blue-600 hover:underline ml-1" target="_blank" rel="noopener noreferrer">Link ↗</a></span>
    </li>
    <li className="flex gap-2" id="source-5">
      <span className="font-semibold text-gray-500">[5]</span>
      <span>Anthropic. "Pricing — Claude API Docs". platform.claude.com, 2026. <a href="https://platform.claude.com/docs/en/about-claude/pricing" className="text-blue-600 hover:underline ml-1" target="_blank" rel="noopener noreferrer">Link ↗</a></span>
    </li>
    <li className="flex gap-2" id="source-6">
      <span className="font-semibold text-gray-500">[6]</span>
      <span>GitHub. "Improving token efficiency in GitHub Agentic Workflows". github.blog, 2026. <a href="https://github.blog/ai-and-ml/github-copilot/improving-token-efficiency-in-github-agentic-workflows/" className="text-blue-600 hover:underline ml-1" target="_blank" rel="noopener noreferrer">Link ↗</a></span>
    </li>
    <li className="flex gap-2" id="source-7">
      <span className="font-semibold text-gray-500">[7]</span>
      <span>Tom's Hardware. "OpenClaw creator burns through $1.3 million in OpenAI API tokens in a single month — 603 billion tokens across 7.6 million requests and 100 coding agents". 2026. <a href="https://www.tomshardware.com/tech-industry/artificial-intelligence/openclaw-creator-burns-through-1-3-million-in-openai-api-tokens-in-a-single-month" className="text-blue-600 hover:underline ml-1" target="_blank" rel="noopener noreferrer">Link ↗</a></span>
    </li>
    <li className="flex gap-2" id="source-8">
      <span className="font-semibold text-gray-500">[8]</span>
      <span>MakeUseOf. "Someone left Claude Code running overnight, and it cost $6,000". makeuseof.com, May 2026. <a href="https://www.makeuseof.com/someone-left-claude-code-running-overnight-and-it-cost-6000/" className="text-blue-600 hover:underline ml-1" target="_blank" rel="noopener noreferrer">Link ↗</a></span>
    </li>
    <li className="flex gap-2" id="source-9">
      <span className="font-semibold text-gray-500">[9]</span>
      <span>Godnick, Henry. "What Happened When I Got a Surprise $80 Claude Bill". dev.to, 2026. <a href="https://dev.to/godnick/what-happened-when-i-got-a-surprise-80-claude-bill-5d83" className="text-blue-600 hover:underline ml-1" target="_blank" rel="noopener noreferrer">Link ↗</a></span>
    </li>
    <li className="flex gap-2" id="source-10">
      <span className="font-semibold text-gray-500">[10]</span>
      <span>Lindenbauer, Tobias et al. (JetBrains Research). "The Complexity Trap: Simple Observation Masking Is as Efficient as LLM Summarization for Agent Context Management". arXiv:2508.21433, August 2025 (v3, October 2025). <a href="https://arxiv.org/abs/2508.21433" className="text-blue-600 hover:underline ml-1" target="_blank" rel="noopener noreferrer">Link ↗</a></span>
    </li>
    <li className="flex gap-2" id="source-11">
      <span className="font-semibold text-gray-500">[11]</span>
      <span>Anthropic. "Manage costs effectively — Claude Code Docs". code.claude.com, checked October 2026. <a href="https://code.claude.com/docs/en/costs" className="text-blue-600 hover:underline ml-1" target="_blank" rel="noopener noreferrer">Link ↗</a></span>
    </li>
    <li className="flex gap-2" id="source-12">
      <span className="font-semibold text-gray-500">[12]</span>
      <span>Anthropic. "Model configuration — Claude Code Docs". code.claude.com, checked October 2026. <a href="https://code.claude.com/docs/en/model-config" className="text-blue-600 hover:underline ml-1" target="_blank" rel="noopener noreferrer">Link ↗</a></span>
    </li>
  </ol>
</div>