@Youssofal_: Thank you to Kate for the comprehensive review of MTPLX. She’s tested multiple different MLX runtimes and concluded MTP…
Summary
nicekate tested multiple MLX runtimes on a Mac for running Qwen3.6-27B and concluded that MTPLX is the fastest, achieving 43 tok/s at 4bit quantization.
View Cached Full Text
Cached at: 05/16/26, 07:21 PM
Thank you to Kate for the comprehensive review of MTPLX.
She’s tested multiple different MLX runtimes and concluded MTPLX is the fastest!
The biggest realisation is I accidentally stripped the vision head from the model. I’ll restore it soon.
I’m looking forward to bringing major upgrades to MTPLX over the weekend.
nicekate (@nicekate8888): 最近二十天我都在折腾一件事——怎么让 Qwen3.6-27B 在我的 Mac 上跑得又快又好。
一开始我用 Unsloth Q5，18 tok/s，风扇呼啦呼啦响。
后来换成 MLX 6bit + DFlash，提到 22 tok/s，还是不够快。
直到我遇到了 MTPLX 4bit，43 tok/s，质量不错。
完整视频：🧵
Similar Articles
@Youssofal_: 72+ TPS on Qwen 3.6 27B on a Macbook pro M5 max. MTPLX V2 out now! The fastest way to run models on MLX.
MTPLX V2 is released, claiming 72+ tokens per second on Qwen 3.6 27B running on a Macbook Pro M5 Max via MLX.
MTPLX V1: The Swift App For Running & Creating MLX MTP Models (2x TPS Qwen 3.6 27B)
MTPLX V1 is a native Mac app that bundles the MTP speculative decoding engine for MLX models, offering features like model conversion via Forge, built-in chat, benchmarking, and support for smaller models. It achieves over 2x speedup with mathematical exactness.
@Youssofal_: MTPLX V0.3 Is Out!: - I realised M1 & M2 macs do not support BF16 and were emulating it leading to significantly decrea…
MTPLX v0.3 is released, a native runtime for Apple Silicon that uses Multi-Token Prediction (MTP) to double decode speed while maintaining distributional accuracy via Leviathan-Chen acceptance.
@nash_su: Mac inference speed doubled. MTPLX is an integrated solution combining MLX and MTP, specifically optimized for model inference on Apple Silicon. By using models with a custom MTP head, it can deliver doubled inference speed. I tested it with Qwen3.6-27…
MTPLX is an integrated solution combining MLX and MTP, specifically optimized for model inference speed on Apple Silicon. Tests show that Qwen3.6-27B achieves double the inference speed of LM Studio, and it also integrates fan management.
@nicekate8888: For the past twenty days, I've been obsessing over one thing — how to make Qwen3.6-27B run fast and well on my Mac. I started with Unsloth Q5, got 18 tok/s, and the fan was roaring. Then I switched to MLX 6bit + DFlash, hitting 22 tok/s, still not fast enough. Eventually I found MTPLX 4bit: 43 tok/s with good quality.
The user shares their experience optimizing Qwen3.6-27B inference speed on a Mac using different quantization methods (Unsloth Q5, MLX 6bit + DFlash, MTPLX 4bit), ultimately reaching 43 tok/s.