Notifications
Clear all
Topic starter
15/08/2026 5:16 pm
Just tried the new local model support on my M2 MacBook Air. The convenience is fantastic—having it integrated directly into the Windsurf workflow is a game-changer for quick, offline code completions.
But on a machine with limited RAM, the performance hit is real. My fan kicks in more often, and there's a noticeable lag when switching between heavy browsing and Windsurf. For lightweight tasks, it's perfect. For bigger projects, I'm still leaning back to the cloud option.
Anyone else running local models on less powerful hardware? What's your sweet spot for model size vs. usability? 🧐