Layr-Labs/qwen-3.8-mtp-challenge

Empirical benchmarks and quality evaluations of Qwen3.8-27B with Multi-Token Prediction (MTP) across Splash, MTPLX, MLX, and llama.cpp on Apple Silicon

qMLX: Custom inference engine for Qwen 3.5 122B on Apple Silicon, extending MLX with hybrid attention support, SSD-backed KV cache, and RYS layer duplication for efficient large-model serving on consumer hardware.

Qwen3.8-Flash-Next on Apple Silicon: an MLX engine that reads Strata's GGUF files and speaks its line protocol (unofficial)

Qwen 3.6 - llama.cpp (with or without MTP) vs mlx-ml benchmark

Fork of antirez/ds4 for Apple Silicon, tuned mostly for M1/M2: Qwen3.8-Flash-Next with 262K context on 64 GB and up, 35 tok/s on an M1 Max, MTP and vision, one-line install.

Community-driven benchmark suite for MLX inference engines on Apple Silicon

Quasar is an inference engine for Qwen3.6-27B-MTP / Qwen3.8-27B on consumer NVIDIA cards as well as Apple Silicon, in the spirit of antirez/ds4

Up to 4× faster LLM decoding on Apple Silicon, lossless. Native MLX port of DeepSeek's DSpark & z-lab's DFlash speculative decoding — Gemma-4, Qwen3.8, Muse-Glimmer, Nemotron, LFM2.5, Ornith-1.0, ternary Bonsai-27B.

Exact speculative decoding on Apple Silicon, powered by MLX.