Comment by meander_water

Just looking through your results, seems like gpt-6 luna was run with reasoning:off for a lot (all?) results. Seems like an unfair comparison.