@zebassembly I was doing this (100USD Kimi plan + 20 USD GPT plan + 20 USD Deepseek). Built a working workflow around. I switched to 200 USD tiers GPT and Claude (400 USD in total, which is more than double) but the output quality is way better and I am losing less time overall on my work.
Full content is available at the original source.
x.com
// related articles