# youssofal/MTPLX

<https://repo.masters-masters.com/en/repos/youssofal/MTPLX>

[Home](https://repo.masters-masters.com/en) › [youssofal](https://repo.masters-masters.com/en/owners/youssofal) › [MTPLX](https://repo.masters-masters.com/en/repos/youssofal/MTPLX)

youssofal/MTPLX has 2,671 stars on GitHub. Its RepoMasters card: RC tier. The fastest way to run Qwen 3.8 Flash Next, Qwen 3.8 27B and Ternary Bonsai 2 27B on…

| Field | Value |
| --- | --- |
| Owner | [youssofal](https://repo.masters-masters.com/en/owners/youssofal) |
| Stars | 2,671 |
| Forks | 209 |
| Language | [Python](https://repo.masters-masters.com/en/languages/python) |
| License | Apache-2.0 |
| Created | 2026 |
| Topics | apple-silicon, local-ai, mlx, mtp, openai-compatible, qwen, speculative-decoding, anthropic-compatible, inference-engine, metal |
| Tier | [RC (2,000 to 4,999 ★)](https://repo.masters-masters.com/en/ranks/blockbuster) |
| View on GitHub | [youssofal/MTPLX](https://github.com/youssofal/MTPLX) |

## Repository

The fastest way to run Qwen 3.8 Flash Next, Qwen 3.8 27B and Ternary Bonsai 2 27B on a Mac: 125 tok/s in OpenCode on an M5 Max, and a 27B model on 16 GB Macs. Native MTP speculative decoding on Apple Silicon, exact at any temperature. OpenAI and Anthropic compatible local server.

The tier for repos with 2,000 to 4,999 ★. Published odds per card pulled: 3%. Rarity: Epic.

## Nearby in Python

- [IceClear/StableSR · ★ 2,671 · Python · RC](https://repo.masters-masters.com/en/repos/IceClear/StableSR)
- [Memento-Teams/Memento · ★ 2,671 · Python · RC](https://repo.masters-masters.com/en/repos/Memento-Teams/Memento)
- [MetapriseAI/OrgKernel · ★ 2,671 · Python · RC](https://repo.masters-masters.com/en/repos/MetapriseAI/OrgKernel)

---
Data: GitHub · Illustration drawn by RepoMasters from the repo’s ID and its language color.