{"id":1766,"date":"2026-02-23T15:37:22","date_gmt":"2026-02-23T23:37:22","guid":{"rendered":"http:\/\/blog.nillsf.com\/?p=1766"},"modified":"2026-02-23T15:37:23","modified_gmt":"2026-02-23T23:37:23","slug":"the-joy-of-switching-ai-models","status":"publish","type":"post","link":"https:\/\/nillsf.com\/index.php\/2026\/02\/23\/the-joy-of-switching-ai-models\/","title":{"rendered":"The joy of switching AI models"},"content":{"rendered":"\n<p class=\"wp-block-paragraph\">I\u2019ve noticed something interesting while building software with AI agents: sometimes the fastest way to fix a stuck coding agent is simply to switch to another model.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">I started building software again using GitHub Copilot CLI. One\nthing I noticed is how each model is slightly different. The models themselves\nkeep improving and keep getting better; but each one almost has like a\npersonality to it.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Not only the models are improving, but the tools are as\nwell. There are people online who went from Cursor to Claude Code and now over\nto OpenAI\u2019s Codex tooling. Even OpenClaw\u2019s creator switched from Claude Code\nover to OpenAI\u2019s Codex recently.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">I haven\u2019t gone that far yet, but one thing I have noticed\nmyself doing is switching between Claude Open 4.6 and OpenAI\u2019s gpt-5.3-codex as\nthe coding model I\u2019m using in Copilot CLI. <\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Why you ask? Because I feel like sometimes the LLMs get\nstuck on a problem, and putting another LLM in the same context window somehow\nbreaks that loop. My typical process boils down to:<\/p>\n\n\n\n<ol class=\"wp-block-list\"><li>Start a plan with opus 4.6. <\/li><li>Start coding with opus 4.6. <\/li><li>If I notice bad results or the agent getting\nstuck, switch to gpt-5.3-codex.<\/li><\/ol>\n\n\n\n<p class=\"wp-block-paragraph\">I don\u2019t know if I\u2019m the only one doing this, but it has\nhelped me in more than one occasion. Recently I hit this while building a\ncustom agent for Teams. Opus 4.6 got stuck generating the Teams app manifest\nand kept giving me incorrect manifest files. I switched to GPT-5.3-Codex with\nthe exact same context, and it produced a correct manifest in a single prompt.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">I have seen more advanced use cases of this online, where people are having <a href=\"https:\/\/x.com\/pierceboggan\/status\/2021094988205969465\">different LLMs do competitive code review<\/a>. I\u2019m not that advanced yet. <\/p>\n\n\n\n<p class=\"wp-block-paragraph\">However, I am m interested to see where this goes next. It\u2019d be interesting for tools like GitHub Copilot CLI to automatically select another LLM in case it notices that it\u2019s stuck on a certain bug and keeps looping around. <\/p>\n\n\n\n<p class=\"wp-block-paragraph\">For now, and for me, one thing stands out: Sometimes the best way to move forward is not to push the same model harder, but to change the mind that is solving the problem.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>I\u2019ve noticed something interesting while building software with AI agents: sometimes the fastest way to fix a stuck coding agent is simply to switch to another model. I started building software again using GitHub Copilot CLI. One thing I noticed is how each model is slightly different. The models themselves keep improving and keep getting [&hellip;]<\/p>\n","protected":false},"author":1,"featured_media":1767,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_jetpack_memberships_contains_paid_content":false,"footnotes":""},"categories":[1],"tags":[],"class_list":["post-1766","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-uncategorized"],"jetpack_featured_media_url":"https:\/\/nillsfblog.blob.core.windows.net\/media\/2026\/02\/Screenshot-2026-02-23-153608.png","jetpack_sharing_enabled":true,"_links":{"self":[{"href":"https:\/\/nillsf.com\/index.php\/wp-json\/wp\/v2\/posts\/1766","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/nillsf.com\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/nillsf.com\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/nillsf.com\/index.php\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/nillsf.com\/index.php\/wp-json\/wp\/v2\/comments?post=1766"}],"version-history":[{"count":1,"href":"https:\/\/nillsf.com\/index.php\/wp-json\/wp\/v2\/posts\/1766\/revisions"}],"predecessor-version":[{"id":1768,"href":"https:\/\/nillsf.com\/index.php\/wp-json\/wp\/v2\/posts\/1766\/revisions\/1768"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/nillsf.com\/index.php\/wp-json\/wp\/v2\/media\/1767"}],"wp:attachment":[{"href":"https:\/\/nillsf.com\/index.php\/wp-json\/wp\/v2\/media?parent=1766"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/nillsf.com\/index.php\/wp-json\/wp\/v2\/categories?post=1766"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/nillsf.com\/index.php\/wp-json\/wp\/v2\/tags?post=1766"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}