And increasingly odd benchmarks with no explanation as to the benchmark or even units. Tweets are like "WOWOWOW an AI first. ChatGPT 5.6o-mini high maxxed plus eclise Martian scored a perfect 5π amp-hour on MIT's ASOICTR-exotic benchmark! Anthropics Fable fell short, being 14.5 nanomarks away."

19

I don’t notice that much difference between the two.

6