3 ms·
I don't think it's necessarily bad to have the benchmark, but the graphs of Gemini and Claude doing worse than o3 did kind of leave a bad taste in my mouth. "O
by gwd 1y ago
I don't think it's necessarily bad to have the benchmark, but the graphs of Gemini and Claude doing worse than o3 did kind of leave a bad taste in my mouth. "Oh look, your models are worse than ours at this very important metric that we just made up! How terrible!"
- progbits 1y agoIt's also barely better, within the error bar on some metrics. So if this is the best they can do on their own benchmark that's pretty pathetic and shows they have no edge anymore.