hey, I’m currently getting enough VRAM to run something in the GLM-5.2 range, but I’m wondering: do we actually have a solid ranking that compares closed-source and open-weight LLMs side by side? I’ve been trying to find a clear “closed vs open” leaderboard, but most benchmarks feel fragmented or don’t really answer the practical question of what’s actually best to run locally versus what’s only competitive through API models. Also, are there any open models that feel as impressive for th...