Qwen3 4B outperforms cloud agents on code tasks—with Mahoraga research [R]
About this article
Hey everyone in ML. I've been working on Mahoraga, an open-source orchestrator that routes tasks across local and cloud AI agents using a contextual bandit (LinUCB) that learns from every decision. Context (skip): I only started integrating AI into my workflows in late 2025, so I came on the scene broke with no credits. This left me with local models. However, many students and employees also receive credits from their institution to work with. (I got claude yippee) I wanted to be able to fla...
You've been blocked by network security.To continue, log in to your Reddit account or use your developer tokenIf you think you've been blocked by mistake, file a ticket below and we'll look into it.Log in File a ticket