6 ms·
Not entirely unexpected given that these models have 2x - 5x more parameters than the 120b gpt-oss model. It’s impressive their performance gets so close given
by davely 1y ago
Not entirely unexpected given that these models have 2x - 5x more parameters than the 120b gpt-oss model.
It’s impressive their performance gets so close given the size!