Hi lovely ladies, Anthropic released Claude Opus 5 on Friday. If you're on Claude Pro it is now the strongest model you can choose and on Max it's already the default. Nothing to buy, nothing to install. Open the model switcher and it's sitting there. 👇🏼 Three things, not one Everyone is repeating the same headline, so here are the two nobody is mentioning. One. It holds a million tokens. That's roughly 750,000 words in a single conversation. Your brand voice guide, a year of captions, every discovery call transcript, your whole offer suite, your analytics export, and the messy Google Doc where you dumped forty ideas at 11pm. All of it at once. Anthropic says it stays consistent right across that window rather than going vague in the middle, which was always the problem with long chats before. Two. It thinks by default now. You no longer have to talk it into reasoning. It decides how hard to think, on its own, every turn. Which quietly moves the work onto us, because it will now reason very carefully about whatever you handed it. Hand it nothing and it will reason very carefully about nothing. Three. There's a dial. Five effort levels, from low up to max. Anthropic says asking for more effort turns into better answers more reliably than it did on any earlier Opus. So "think properly about this before you answer" has stopped being a placebo. It's a real setting, and in the app you reach it just by saying so. 😬 The bit I keep seeing Most of us are still using AI the way we used it in 2023. Open a chat. Type "give me ten hooks for my coaching offer". Get ten hooks written for nobody in particular. Feel a bit flat about it. Close the tab. Decide it doesn't get your voice. I have done this more times than I'd like to admit. 🤪 Of course it doesn't get your voice. You never gave it one. It's like asking someone you have never met to write your wedding speech, then being annoyed they got your sister's name wrong. The gap opening up right now isn't ChatGPT versus Claude. Both will keep improving. The gap is between the people who feed the thing properly and the people who keep asking it to guess.