Claude that I've been using sparingly
It was originally reset on the 8th, but the date suddenly changed and the usage reset, yet Codex gets used up so quickly
This is Ollama Cloud whose usage reset at 9 AM this morning
I haven't done much, but the token usage is draining fast
For Claude I use Sonnet 4.6, for Codex I use 5.4 mini, and for Ollama Cloud I use about glm-5
Lately it seems like since DeepSeek kicked things off with price cuts first, other companies are also lowering token costs or increasing the allowance for flat-rate plans — please increase it faster, I'm getting dizzy here
I tried using gemma4:12b and on a device with an AMD AI Max 395, it gets about 23 tokens
It seems usable locally, but it feels way too slow — you could set it to work and go hang out and come back... yo
