Grok 2 and Llama 3.1 don't stand out from the pack

Both models work fine, but aren't noticeably better than existing leaders.

Today Twitter (aka X) finally gave me access to Grok 2, the language model announced by Elon Musk’s AI company last week. I also haven’t yet published anything about the performance of the largest Llama 3.1 model, which Meta released last month.

So I thought I’d do a quick post comparing their performance on some of the brain teasers I’ve used in the past