Qumeric comments on Recent AI model progress feels mostly like bullshit

[Qumeric]

I am curious to see what would be the results of the new Gemini 2.5 pro on internal benchmarks.