<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Pricing on Kuldeep Pisda</title><link>https://kdpisda.in/tag/pricing/</link><description>Recent content in Pricing on Kuldeep Pisda</description><generator>Hugo</generator><language>en-us</language><lastBuildDate>Fri, 25 Sep 2026 09:00:00 +0530</lastBuildDate><atom:link href="https://kdpisda.in/tag/pricing/index.xml" rel="self" type="application/rss+xml"/><item><title>Claude's Refusals Have Been Free Since June. Three Categories Just Stopped Being Free</title><link>https://kdpisda.in/anthropic-refusal-billing-bio-frontier-llm-reasoning-extraction/</link><pubDate>Fri, 25 Sep 2026 09:00:00 +0530</pubDate><guid>https://kdpisda.in/anthropic-refusal-billing-bio-frontier-llm-reasoning-extraction/</guid><description>&lt;p&gt;Since June 2, a Claude API request that gets refused before the model produces any output has cost nothing. You get back a normal HTTP 200, &lt;code&gt;stop_reason: &amp;quot;refusal&amp;quot;&lt;/code&gt;, &lt;code&gt;usage.output_tokens: 0&lt;/code&gt;, and no line item. On September 24, Anthropic carved three specific exceptions into that rule, documented in one paragraph on its release notes page: refusals whose &lt;code&gt;stop_details.category&lt;/code&gt; is &lt;code&gt;&amp;quot;bio&amp;quot;&lt;/code&gt;, &lt;code&gt;&amp;quot;frontier_llm&amp;quot;&lt;/code&gt;, or &lt;code&gt;&amp;quot;reasoning_extraction&amp;quot;&lt;/code&gt; are now billed like any other request, at the full rate of the model that ran, even though it wrote zero tokens.[1] If you have retry logic, an evaluation harness, or Anthropic&amp;rsquo;s own server-side fallback pointed at Claude, this is a real cost change on traffic that looks identical to what it looked like a week ago.&lt;/p&gt;</description></item></channel></rss>