whichllm - 내 하드웨어에서 실제로 돌아가고 최고 성능을 내는 로컬 LLM 찾기

whichllm solves the critical challenge of selecting optimal local LLM models by leveraging actual benchmark performance across multiple sources rather than parameter counts alone. By intelligently accounting for hardware constraints (VRAM, architecture, bandwidth) across NVIDIA, AMD, and Apple Silicon GPUs, it removes guesswork from model selection. The tool further enables hardware planning simulations, allowing users to validate GPU upgrade decisions before purchase.