[UrineSqueegee]

I'd really wait for additional neutral benchmarks, I asked the 20b model on low reasoning effort which number is larger 9.9 or 9.11 and it got it wrong.

Qwen-0.6b gets it right.