<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Model Selection on Notes from the Rabbit Hole</title><link>https://magnus919.com/tags/model-selection/</link><description>Recent content in Model Selection on Notes from the Rabbit Hole</description><generator>Hugo</generator><language>en</language><copyright>© [Magnus Hedemark](https://github.com/magnus919)</copyright><lastBuildDate>Sun, 02 Aug 2026 00:31:37 -0400</lastBuildDate><atom:link href="https://magnus919.com/tags/model-selection/index.xml" rel="self" type="application/rss+xml"/><item><title>GPT-5.6 Luna, Terra, Sol, and DeepSeek V4 Flash: The Cost of Finished Work</title><link>https://magnus919.com/2026/08/stop-using-your-middle-model-by-default/</link><pubDate>Sun, 02 Aug 2026 00:31:37 -0400</pubDate><guid>https://magnus919.com/2026/08/stop-using-your-middle-model-by-default/</guid><description>&lt;p>The cheapest AI model isn&amp;rsquo;t the one with the lowest token price. It&amp;rsquo;s the one that gets acceptable work across the finish line with the fewest retries, repairs, and wasted minutes.&lt;/p>
&lt;p>That sounds obvious, but it is not how most of us pick models. When a model family has a cheap tier, an expensive tier, and a middle tier, the middle feels responsible. It sounds like the sensible default.&lt;/p>
&lt;p>OpenAI&amp;rsquo;s &lt;a href="https://openai.com/index/advancing-the-price-performance-frontier-with-gpt-5-6/">GPT-5.6 family&lt;/a> makes this temptation easy to see. Sol is the frontier model. Terra is the middle. Luna is the inexpensive, high-volume option. A reasonable first instinct is to use Terra most of the time, move up to Sol when the work looks hard, and drop to Luna when it looks trivial.&lt;/p></description></item></channel></rss>