<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Openai on Start AI Tools - Presented by Intent Solutions</title><link>https://startaitools.com/tags/openai/</link><description>Recent content in Openai on Start AI Tools - Presented by Intent Solutions</description><generator>Hugo</generator><language>en-US</language><copyright>Intent Solutions. All rights reserved.</copyright><lastBuildDate>Sat, 04 Jul 2026 07:15:00 -0600</lastBuildDate><atom:link href="https://startaitools.com/tags/openai/index.xml" rel="self" type="application/rss+xml"/><item><title>Shipping gpt-5.4 as One Config Line</title><link>https://startaitools.com/posts/eval-gated-model-swap-gpt-5-4/</link><pubDate>Fri, 03 Jul 2026 10:00:00 -0600</pubDate><guid>https://startaitools.com/posts/eval-gated-model-swap-gpt-5-4/</guid><description>&lt;p&gt;On 2026-07-03, DiagnosticPro shipped gpt-5.4 in production. The structural finding: gpt-4o silently capped reports at ~1,225 words across 36 benchmark runs—no matter the prompt. gpt-5.4 delivers depth. A real stored report came back at 15 sections, 3,393 words.&lt;/p&gt;
&lt;p&gt;The rollout was step 2 of a two-axis, independently eval-gated ladder:&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Step 1 (June):&lt;/strong&gt; Prompt framework v3.0 (v3-b few-shot exemplar) beat v2.0 18/18 blind. Later corroborated by adding a third model to the grid.&lt;/p&gt;</description></item></channel></rss>