mlx has 1.1M monthly downloads and 27K GitHub stars. Version 0.31.2 supports Python 3.10+, first published in 2023.
It is actively maintained under a permissive license, last released in Apr 2026.
It is built on top of mlx-metal. It is used by 301 tracked packages, including mlx-lm, unsloth-zoo and mlx-vlm.
GitHub topics include mlx.
Package Insights
((week_daily_avg - month_daily_avg) / month_daily_avg) * 100Weekly Downloads
GitHub Stars
Downloads by OS
Python Versions
Top Countries
Dependencies
- mlx-metal ==0.31.2
10 optional dependencies
- ml-dtypes[dev]
- mlx-cpu[cpu]
- mlx-cuda-12[cuda]
- mlx-cuda-12[cuda12]
- mlx-cuda-13[cuda13]
- numpy[dev]
- pre-commit[dev]
- psutil[dev]
- torch[dev]
- typing-extensions[dev]
Used By
- mlx-lm
- unsloth-zoo
- mlx-vlm
- mlx-audio
- tabpfn
- mlx-whisper
- mlx-embeddings
- rapid-mlx
- mflux
- parakeet-mlx
- mlx-openai-server
- mlx-optiq
- vllm-mlx
- mlx-tune
- vmlx
- mlx-gen
- ezmsg-sigproc
- squish-ai
- veloxquant-mlx
- edge-studio
- mlx-qwen3-asr
- mtplx
- f5-tts-mlx
- instructlab
- lucid-dl
- mlx-taef
- mlx-lm-lora
- mlx-arsenal
- mlx-mfa
- leann-core
- vocos-mlx
- lightning-whisper-mlx
- mlx-memo
- mlx-whisperx
- mlx-teacache
- mlx-embedding-models
- moshi-mlx
- mlx-video-with-audio
- speech-to-speech
- whisper-turbo-mlx
- mlx-lattice
- turboquant-mlx-full
- pymss
- diffusionkit
- mlx-embeddings-lora
- mlx-speech
- mlx-kquant
- whisper-mps
- mlx-omni-server
- atai-whisper-tool