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
Every time these types of posts are discussed at work, the point is always that there's more risk of falling behind (more like FOMO) at pace if others are able to launch or bring new features faster