Comment by leumon 15 hours ago So 89.4% on Terminal Bench 2 but only 19.1% on Tbench 4. Opus 5 is 89.1%/51.8%. 2 comments leumon Reply aszen 11 hours ago I was thinking the same, obvious suspicion is they benchmaxed it on older bench. the_duke 42 minutes ago Google has always done quite a lot of benchmaxing for Gemini.
aszen 11 hours ago I was thinking the same, obvious suspicion is they benchmaxed it on older bench. the_duke 42 minutes ago Google has always done quite a lot of benchmaxing for Gemini.
I was thinking the same, obvious suspicion is they benchmaxed it on older bench.
Google has always done quite a lot of benchmaxing for Gemini.