I have only a Codex Pro subscription. Yes, I exhaust the limits regularly. Then, I go for OpenCode Zen and just pay until the limits refresh.
I think they’re very capable. I’d say that the Anthropic Models are better, however, very often this level of intelligence is simply not required. The OpenAI models definitely give you more for the price. I mean, Luna is almost for free, bit still does a great job for exploration or planning.
I usually use Luna on XHigh and Sol on Low or Medium
are they good? do you sometimes exhaust your token limits?
commit 8f36395