# youssofal/MTPLX

3x faster speeds on MLX | Qwen 3.8 27B | Native MTP Speculative Decoding On Apple Silicon With No External Drafter.

Repository: https://github.com/youssofal/MTPLX
Canonical: https://ross.abutalabs.com/products/mtplx
Homepage: https://mtplx.com
Language: Python
License: Apache-2.0
License Family: permissive
Topics: apple-silicon, local-ai, mlx, mtp, openai-compatible, qwen, speculative-decoding, anthropic-compatible, inference-engine, metal, qwen3-next, speculative-sampling, mtplx, native-mtp
Last push: 2026-08-26T00:45:59+00:00

## Health v2 (maintenance only)
Score: 81/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 99, release rhythm 99, longevity 8
- inputs: {"age_days": 124, "days_push": 8, "days_rel": 8, "gap_med": 0.0, "n_releases_24m": 43}
- flags: young
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1698, forks 127 (observed 2026-08-28T04:05:22.411301+00:00)

## What it is
MTPLX is a native macOS app and CLI that runs local LLMs on Apple Silicon roughly twice as fast using native multi-token prediction (MTP) speculative decoding, with exact rejection sampling that preserves output distribution. It ships an OpenAI/Anthropic-compatible local server, a built-in chat UI, and a model forge that converts Hugging Face models to MLX.

## Use cases
- run local llms faster on my mac
- serve an openai-compatible local model endpoint on apple silicon
- speed up qwen inference with speculative decoding
- chat with local models offline on m-series macs
- convert a hugging face model to mlx and measure speedup
- power coding agents like opencode with a local model
- run llm inference without an external draft model

## When to choose
- you have an Apple Silicon Mac (M1+, macOS 14+) and want faster local LLM inference with identical output
- you use MTP-capable models like Qwen 3.5/3.6/3.8 and want native speculative decoding without a second draft model
- you need an OpenAI- or Anthropic-compatible local server for coding agents and tools
- you want a one-click Mac app experience including model download and hardware-based recommendations

## When to avoid
- you run on Linux, Windows, or non-Apple hardware
- you need models without MTP heads or broad model-family support
- you need multi-GPU or datacenter-scale serving rather than single-machine local inference
- you require a runtime with a long track record or large maintainer team

## Facets
- artifact type: application
- maturity: active
- function: llm-inference, cli, chatbot
- domain: large-language-models, machine-learning, self-hosted, developer-tools
- platform: cli
- tags: speculative-decoding, multi-token-prediction, mlx, apple-silicon, openai-compatible, anthropic-compatible, local-llm, qwen, metal, inference-engine, macos, apple-ecosystem

## Member repositories
- youssofal/MTPLX (main) score 81

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:22.411301+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T03:37:50.295409+00:00, confidence not recorded.
  - readme: https://github.com/youssofal/MTPLX (fetched 2026-08-28T04:05:22.411301+00:00, sha 9e48cb2500d7)
  - homepage: https://mtplx.com (fetched 2026-08-29T11:12:39.399726+00:00, sha 12524fe1f9c4)
  - registry_pypi: https://pypi.org/pypi/mtplx/json (fetched 2026-08-29T11:12:39.413179+00:00, sha 77d01c735c27)
  - site_page: https://mtplx.com/faq (fetched 2026-08-29T11:12:39.408685+00:00, sha e8935b0f4d53)
  - site_page: https://mtplx.com/releases (fetched 2026-08-29T11:12:39.410733+00:00, sha 74e246e30bb8)
- Data as of 2026-08-30T08:39:29.467469+00:00.