Apps & Tools
MTPLX
3x faster Qwen 3.8 27B on Mac via native MTP and DFlash2 speculative decoding
bydavidtai
Open source
MTPLX media is blocked
Allow external media to connect to the provider and play this content.
Description
A high-performance implementation of Qwen 3.8 27B optimized for Apple Silicon using the MLX framework. It features native Multi-Token Prediction (MTP) and a pinned DFlash2 drafter to achieve significantly higher decode throughput without requiring an external drafter model.
Descriptions, tags, and model credits may be AI-generated or inferred from public sources and can be incomplete or wrong. Learn more.