AI coding agents are proliferating, but the economics of running large language models (LLMs) are breaking down as developers juggle multiple APIs and seriously unpredictable token bills. This is particularly problematic when agents start making dozens of model calls just to complete a single request.
The response from the developer community has been open-source coding agents, which operate one layer above the models. They keep costs consistent because they’re independent of — and work across — many LLMs.
OpenCode is one such player,…
