mlx-vlm

LLM inference server with continuous batching & SSD caching for Apple Silicon — managed from the macOS menu bar

OpenAI and Anthropic compatible server for Apple Silicon. Run LLMs and vision-language models (Llama, Qwen-VL, LLaVA) with continuous batching, MCP tool calling, and multimodal support. Native MLX backend, 400+ tok/s. Works with Claude Code.

Fully Open Framework for Democratized Multimodal Training

Open-source evaluation toolkit of large multi-modality models (LMMs), support 220+ LMMs, 80+ benchmarks