Anthropic just released Claude Opus 5.5, and from what I've seen so far, it's better.
That's a big deal for me, because the last version drove me nuts. I'd ask Opus 5 for a document and get British spellings everywhere. It would forget the brand rules I'd just given it.
Claude Opus 5.5 is out, and early signs say Anthropic fixed the things that frustrated Claude users most.
Neil used the previous Opus 5 a lot. It mixed in British spellings, forgot brand rules it had just been given, and sometimes got stuck in thinking loops that ate through
Claude Sonnet 5.5 beats Opus 5.5 in the terminal work test.
It's half the price. 1️⃣ Sonnet 5.5 approaches Opus 2️⃣ GPT-6.1 Sol is 1/5th of Astra class 3️⃣ Improve your Japanese with ElevenLabs v4 Today's axis is a medium model. 👇Continue in the video (4 minutes)
If you think Sonnet 5.5 made the horse’s running better than Sonnet 5, I’ve got some bad news for you
But compared to Opus 5.5, Sonnet is clearly the winner here
You have reached the end of the archive
All of Claude Opus 5