# youssofal/MTPLX

> Run local LLMs on Apple Silicon with native multi-token prediction (MTP) speculative decoding and no external drafter; speeds up inference 1.6–2.2x while keeping the same output distribution.

- Magnitude: 1.1 out of 10 — Steady
- Stars: 2,410 total · +1 star on 2026-09-23 (UTC), ≈ 4 by evening
- Star trust: star growth looks organic
- Category: Language models · Language: Python · License: Apache-2.0 · Created: 2026-05-02 · Last push: 2026-09-19
- GitHub: https://github.com/youssofal/MTPLX · Homepage: https://mtplx.com · Page: https://gitnova.dev/en/r/youssofal/MTPLX

## Useful for

- Start a local OpenAI-compatible server on a Mac and point Claude Code or Open WebUI at it
- Auto-tune MTP draft depth for a specific chip with mtplx tune
- Build and verify an MTP adapter for a Hugging Face model via mtplx forge

## Why it’s here

- 1 star so far today, about 4 expected by the end of the day.

## Star trust

Star growth looks organic.

Star-trust labels are heuristics based on the repository’s behavior, not a check of every stargazer.

## Numbers

- Forks: 185
- Issues and pull requests: 518
- Watchers: 20
- Average over the last week: 11 per day
- Usual pace: 28 per day
- Latest release: v2.11.3 (2026-09-17)

## Stars per day, last 30 days (oldest → newest, today is partial)

2026-09-30 … 2026-10-29: None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None

## Magnitude by day, last 11 days

09-13 2.2, 09-14 1.8, 09-15 1.8, 09-16 1.7, 09-17 1.7, 09-18 2.1, 09-19 1.9, 09-20 1.6, 09-21 1.7, 09-22 1.6, 09-23 1.4

---

Magnitude (0–10) measures how fast and how unusually interest in a repository is growing right now. It is not a quality score.
Days are UTC. “So far today” is a fact; “expected by the end of the day” is a forecast.
Summaries and use cases are written by an LLM (DeepSeek V4.1 Flash) from the README and may be inaccurate: verify specific claims (benchmarks, speed, hardware) in the repository itself.
Data as of 2026-09-23 10:35 UTC, updated every 30 minutes.