Still settling in to what AI models I should be using. Last month, I was happy with GPT-5.6 Sol (xhigh) and fast mode all the time. Astra blew that up because it’s just more expensive. I think GPT-6 Sol (high) is the right balance for my usage.

@manton I am on GPT-6 Sol (high). To be frank, I think less and less about models these days. When things don't go the way I want, I change them.

@manton I usually plan and review in Sol medium or high, implement in Terra medium or high for my day job. For personal, I’m a little more frugal but I use that as my own training to understand token efficiency.

Hopefully the 6 series makes me break out xhigh even less when I feel like I’m in review finding circles.