Qwen3.6 35B A3B 4bit MTPLX Optimized Speed This is an MLX 4 bit build of Qwen/Qwen3.6 35B A3B packaged for fast local serving with lightning mlx . The model includes an MTPLX sidecar ( mtp.safetensors ) and runtime metadata ( mtplx runtime.json ) so lightning mlx can use its Qwen3.6 MTPLX serving path on Apple Silicon. The included runtime metadata was verified on Darwin arm64 with mtplx version: 0.1.0rc3 , mtp depth max: 1 , and the performance cold recommended profile. Refer to the original Qwen3.6 35B A3B model card for base model capabilities, license, and upstream details. Install lightning mlx Install directly from GitHub: Or use the self contained installer: Verify the CLI: Serve this model Serve directly from Hugging Face: Or serve from a local checkout: For long running local use, start it as a daemon: Daemon mode starts a detached supervisor, writes logs under ~/.lightning mlx/logs/ , and can restart the server if the model process exits unexpectedly. Useful daemon commands: Use status to list running daemons, tui to attach the live monitor, and kill to stop by supervisor PID, server PID, alias, or model name. Use the OpenAI compatible API Once the server is running, send…
We use cookies for essential functionality and analytics. You can accept or reject analytics cookies.Cookie policy