Same, i feel that V4 Flash is great at task implementation, but im still looking at bigger models for design. Now, GLM 5.2 with high thinking is actually gettin

Same, i feel that V4 Flash is great at task implementation, but im still looking at bigger models for design. Now, GLM 5.2 with high thinking is actually getting really close now. I have switched for all personal projects right now and am quite happy with the results. I think the magic is in the big context window (1m) + a lot of thinking gets us very close to at least Opus 4.6 level. Im currently running directly on z.ai with a lite coding plan, and have bought API credit on deekseek as well. I will be looking at EU based hosts next and then i might switch over some of the more critical flows.