LM Studio
A desktop application over llama.cpp and MLX, with model discovery and hardware-aware suggestions built in.
From the file· capabilities, not benchmarks
What it is
A graphical client that browses HuggingFace, downloads models, and runs them with a chat interface and a local server. It handles both GGUF and, on Apple Silicon, MLX.
Capabilities
| Model formats | GGUF, MLX |
| KV cache quantization | yes |
| CPU offload | yes |
| MoE expert offload | not supported |
| Multi-GPU | Basic; exposes fewer controls than the engine underneath. |
| Concurrency | Local server for personal use. |
| Platforms | macOS, Windows, Linux |
Best for
People who would rather not use a terminal, and Apple Silicon users who want MLX without setting it up themselves.
Watch out for
Closed source, and its convenience layer hides some of the memory controls that matter when a model is close to not fitting.