<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Multi-Agent on Ahmed</title><link>https://ahmd.io/tags/multi-agent/</link><description>Recent content in Multi-Agent on Ahmed</description><generator>Hugo</generator><language>en-us</language><lastBuildDate>Tue, 02 Jun 2026 00:00:00 +0000</lastBuildDate><atom:link href="https://ahmd.io/tags/multi-agent/index.xml" rel="self" type="application/rss+xml"/><item><title>4.88 Billion Tokens for $120 — Why I Stopped Using Frontier Models for Everything</title><link>https://ahmd.io/blog/2026/06/02/llm-cost-cascading/</link><pubDate>Tue, 02 Jun 2026 00:00:00 +0000</pubDate><guid>https://ahmd.io/blog/2026/06/02/llm-cost-cascading/</guid><description>&lt;p&gt;Last month, I processed &lt;strong&gt;4.88 billion tokens&lt;/strong&gt; through an LLM. My bill: &lt;strong&gt;$120&lt;/strong&gt;.&lt;/p&gt;
&lt;p&gt;If I had run the same workload on a famous frontier model, the minimum estimate would have been &lt;strong&gt;$4,900&lt;/strong&gt;. On other frontier providers&amp;rsquo; cheapest tiers: &lt;strong&gt;$4,880&lt;/strong&gt;.&lt;/p&gt;
&lt;p&gt;I didn&amp;rsquo;t save money by being cheap. I saved money by being intentional about which model does what.&lt;/p&gt;
&lt;hr&gt;
&lt;h2 class="heading" id="the-obvious-thing-nobody-does"&gt;
 The Obvious Thing Nobody Does
 &lt;a href="#the-obvious-thing-nobody-does"&gt;#&lt;/a&gt;
&lt;/h2&gt;
&lt;p&gt;Every major LLM provider markets their flagship model as the default. But here&amp;rsquo;s what the research actually says.&lt;/p&gt;</description></item></channel></rss>