I'm a compsci student and I've been using the 10$ copilot plan for about 2 years now, and it was fine for me since I did a good model distribution taking into account the complexity of the task, I was able to get through the month always using about 80-90% of the plan. But with the addition of 5hour limits, session limits and weekly limits it's almost unusable.
From my research these are the best options for my needs:
- Codex Pro (would like to hear if the limits are that much better compared to Claude or copilot)
- opencode Go (don't really know if the available models would do the trick: qwen 3.5&3.6 plus, deep seek V4 pro, GLM-5.1 (only 880 requests for 5 hours - probably enough idk))
- Kilo + GML (states that has a very good usage limit)
- cursor
- windsurf (people say it decreased it's offering quality recently)
Open to any suggestion
Opencode Go offering seems pretty nice imo, but would like to hear on the usage limits, saw some users say it has a big 5hour limit, but a not so good weekly limit (like if you go 3x through the 5hour limi, the week limit is reached). Would like to know if that's true.