I left Claude for months. Opus 5.5 is why I'm back

Anthropic shipped Claude Opus 5.5, which is 40% cheaper than Opus 5, faster, and built on what Anthropic calls a fundamentally different alignment approach, according to a first-person review by ChatPRD founder Claire Vo. Vo, who had moved most of her daily work to OpenAI's Codex after finding Claude's rambling and hedging annoying, tested Opus 5.5 for a week across four long-running agentic tasks, a ChatPRD homepage redesign, an SVG benchmark, and one firm refusal, and now uses it as her go-to for frontend prototyping while still splitting her model stack with Codex.

Listen or watch on YouTube https://youtu.be/zObYdmNB2Bo , Spotify https://open.spotify.com/episode/04TcTW0pjDliDoqJH71d4p , or Apple Podcasts https://podcasts.apple.com/us/podcast/i-left-claude-for-months-opus-5-5-is-why-im-back/id1809663079?i=1000791159677 I’ve been off Claude for months. Not because it got dumb, but because it got annoying. The rambling, the hedging, the preachy little disclaimers on tasks that didn’t need them. I moved most of my daily work to Codex and I didn’t miss it. Then Anthropic shipped Opus 5.5: 40% cheaper than Opus 5, faster, and with what they’re calling a fundamentally different alignment approach. I ran it for a week across real work, including four long-running agentic tasks, a full ChatPRD homepage redesign, an SVG benchmark, and one very firm refusal, and I’m ready to give you the honest verdict. There’s a lot to like. There are still two things that drive me a little crazy. And there’s one capability I genuinely wasn’t expecting. What you’ll learn: 1. Why I walked away from Claude entirely, and what it took for me to come back 2. The real cost math on Opus 5.5 and why pricing matters more for agentic work than single prompts 3. What happened when I ran four long-running agentic tasks, including one that tried to manipulate Claude mid-run 4. Why Opus 5.5 is now my go-to for frontend prototyping, and where it still lets me down 5. The one capability I genuinely didn’t see coming, and no other model in my stack can match it 6. The moment Opus 5.5 told me flat-out no, and what that says about where Anthropic’s safety posture actually lands in practice 7. Where Codex still wins, and how I’m splitting my model stack after a full week of testing In this episode: 00:00 https://www.youtube.com/watch?v=zObYdmNB2Bo Why I stopped using Claude 01:02 https://www.youtube.com/watch?v=zObYdmNB2Bo&t=62s&pp=0gcJCWMAwfN6Pr3D What Anthropic says Opus 5.5 is 01:54 https://www.youtube.com/watch?v=zObYdmNB2Bo&t=114s Cost, speed, and benchmark overview 03:20 https://www.youtube.com/watch?v=zObYdmNB2Bo&t=200s Safety, alignment, and the cybersecurity limits 05:02 https://www.youtube.com/watch?v=zObYdmNB2Bo&t=302s&pp=0gcJCWMAwfN6Pr3D How I AI bench 05:39 https://www.youtube.com/watch?v=zObYdmNB2Bo&t=339s Voice test: is it actually not annoying? 07:54 https://www.youtube.com/watch?v=zObYdmNB2Bo&t=474s Long-running agentic task results 10:50 https://www.youtube.com/watch?v=zObYdmNB2Bo&t=650s Frontend prototyping 17:23 https://www.youtube.com/watch?v=zObYdmNB2Bo&t=1043s Writing voice and email 19:41 https://www.youtube.com/watch?v=zObYdmNB2Bo&t=1181s SVG illustrations 20:46 https://www.youtube.com/watch?v=zObYdmNB2Bo&t=1246s Video editing 21:42 https://www.youtube.com/watch?v=zObYdmNB2Bo&t=1302s My verdict: what it’s good at, what it still isn’t Tools referenced: • Claude Opus 5.5: https://www.anthropic.com/claude-opus-5-5 https://www.anthropic.com/claude-opus-5-5 • ElevenLabs MCP connector: https://elevenlabs.io/mcp https://elevenlabs.io/mcp • Codex OpenAI : https://openai.com/codex https://openai.com/codex Where to find Claire Vo: ChatPRD: https://www.chatprd.ai/ https://www.chatprd.ai/ Website: https://clairevo.com/ https://clairevo.com/ LinkedIn: https://www.linkedin.com/in/clairevo/ https://www.linkedin.com/in/clairevo/ Production and marketing by https://penname.co/ https://penname.co/ . For inquiries about sponsoring the podcast, email \ email protected\ https://www.lennysnewsletter.com/cdn-cgi/l/email-protection .