I’m not getting my hopes up for this one. Scout and Maverick have been appalling in real world use when compared with models that they claim to outperform on benchmarks.

I’m getting less interested in the benchmarks with each new release since the companies just seem to optimise for them. Every major AI company shoots to near the top of the benchmark with each new model they put out but I don’t often end up switching to those new models since whatever I’m using will magically catch up a week later.

Gemini is poopcrap, not sure why it’s in any comparison matrix