2026.07.11 15:10

Grok 4.5 has been very fast and accurate in most of my test. With gpt 5.5 and claude opus I used to send out a prompt and move on some other task, but now i have to keep coming back to grok 4.5 session cause they keep finishing fast! It’s very interesting how my workflow changed with a new fleet of models and agents. Also updated lfg to support grok 4.5 with grokbuild and cursor And migrating some of the agents to use Grok 4.5 Very exciting times! lfg is an open source multi agent harness that deploy to your vps / local mac with minimal setup. And give you really nice mobile setup like this.

显示更多