Context I have been using OpenCode for about a year. When I started, the available models were things like big pickle, mimo v2.5 and some other weak models. In that time I built more than 10 software tools, all with OpenCode. Same workflow...
Context
I have been using OpenCode for about a year. When I started, the available models were things like big pickle, mimo v2.5 and some other weak models.
In that time I built more than 10 software tools, all with OpenCode. Same workflow, same type of large and complex projects. I never hit my free limit before, it always felt generous.
What is happening now
For the last week or so, I hit the limit very fast. I am using it the same way as before, same project size, same workflow, but now it runs out in under an hour.
Right now I am on the free plan and I use frontier models, mostly DeepSeek-V4.1 Flash, GLM-5.3 Flash or Mimo-v2.6 Flash. The project I am working on is large and complex, but it has been large for a long time(I mean this isn't a new thing). That is why this feels strange, it happened so sudden.
My setup is an i3 4th gen with 8GB RAM, so running a local model is not an option for me.
What I need help with
Has anyone else seen this change recently? Did something change with how free limits are counted, or how context is loaded?
And the main question: how can I get more credits or make the free quota last longer when I do not have a single cent to pay for a plan or API key?
I want to keep building, but right now I get stuck after an hour.
Thanks for reading.