[humanymous]

Opus with low reasoning effort would be faster than Sonnet with high reasoning. So that won't exactly help. I think it would just be what those models are optimized to perform