Is this an official release from the Laya authors?

No, it says so itself. This is an independent MLX port, not an official Convai Innovations release. It provides inference and conversion only; RLCD training and fine-tuning remain in the upstream project.

Will it give me the same answers as the PyTorch model?

On the author's validation fixtures, yes. All three checkpoints matched upstream's selected answer on 63 of 63 validation questions in both FP32 and FP16, 378 of 378 comparisons. That measures fidelity on those fixtures rather than accuracy on arbitrary questions.

What is the catch with that 13.4 ms median?

It is one short English question, P50, on an M3 Max, with model loading excluded. Timing does cover prompt preparation, tokenization, inference, calibration and formatting. Longer inputs, more questions and other hardware all move the number.

Can this run on Linux, or on a machine without an M3 chip?

No. It requires Apple Silicon, Python 3.11 or newer and macOS 14 or later, and the published measurements come from an M3 Max running macOS 27.2. Older supported macOS versions were not tested on that machine, so expect different figures.

Which checkpoints does Laya-MLX support and where do they come from?

The same three upstream checkpoints (laya, laya-multilingual, laya-typed-decisions) with their original weights, prompt formatting, calibration and output schema; pre-converted FP16 versions are on the aac6fef Hugging Face org.

What options speed up repeated Laya-MLX workloads?

Opt into compile=True, pad_to_multiple=16 and cache_prompts=True; the prefix cache is bounded to 128 questions, compilation has a first-use cost, and all three options default to disabled.

What numeric precision does Laya-MLX use by default?

FP16 by default, with float32 available for closer numerical agreement; probabilities can differ slightly across precisions even when the selected label agrees.

CommunityJev Models & Reimplementations

Laya-MLX

A native MLX runtime for Laya's open typed-decision weights: short decisions on Apple Silicon, with no generated tokens, PyTorch, or cloud API.

See highlights

Open github.com

Stars: 6.9k
Language: Python
License: APACHE-2.0
Published by: Community
Author: mizorewww
Last verified: Oct 2026

mizorewww/laya-mlx

User repository on GitHub

View on GitHub

Native MLX runtime for Laya typed decision models — 7–14 ms short decisions on M3 Max. No text generation, PyTorch, or cloud API.

Forks: 553
Last pushed: Updated Oct 2026

Repo stats from the GitHub API, cached Oct 2026.Homepage

About this project

Open-weight typed decisions, running natively on Apple Silicon. 13.4 ms median end-to-end for a short English typed decision. 7.4 ms with the multilingual checkpoint. 0 output tokens. Local MLX inference, with no PyTorch, Transformers runtime, or cloud API. 中文 · Benchmarks · Snake demo · Hugging Face weights.

Quoted from the project's README. Read the full README.

Tagscommunity python open-models local

Works forJev for Classification Jev for Realtime control

Used inJev for Software & DevTools

Languages: Python
Platforms: macOS
Capabilities: Open sourceRuns locally

Highlights

Apache-2.0. 6,730 stars on 3 October 2026. PyPI package laya-mlx.
The README reports a 13.4 ms median for a short English decision, and 7.4 ms on the multilingual checkpoint.
A Snake demo calls Laya on every move. A safety layer can reject an unsafe proposal.
Separate from NandhaKishorM/laya, which is the model, and from laya-coreml, which uses Core ML.

Watch out

Latency figures are the author's, measured on an M3 Max. Not affiliated with TypeSafe or Convai.

Questions about this entry

Short answers drawn from the linked project or page.

Reactions & coverage

Posts, threads, and videos about this entry from around the web.

LinkedIn: What multimodal decision models actually change, and what they do not.

LinkedInJevDirectory.org

Hacker News: Decision Index 0.2.1 scores open reproductions of Jev on a frozen suite.

jackbravo4 pointsDecision Index 0.2.1 – 70 open reproductions of TypeSafe Jev's Decision ModelRead the thread on Hacker News

Hacker News: Pointer to a DiffusionGemma plus vLLM tutorial for a Jev-shaped classifier.

1 pointHow to build a Jev-style classifier with DiffusionGemma and vLLMRead the thread on Hacker News

Hacker News: A one-token logprob wrapper that also classifies webcam frames.

allanrbo151 pointsA single function Jev-like wrapper for LLMs, including vision modelsRead the thread on Hacker News

Related terms

Glossary definitions related to this entry.

Open weights

More Jev resources like this

Laya-CoreML

1.6k

Laya typed decisions through Apple Core ML and the Neural Engine, with the author's speed and energy benchmarks for short decisions on an M3 Max.

Jev Models & Reimplementations#community#python#open-models

Community

mizorewww/laya-coremlPython

OpenDecision

An Apache-2.0 decision engine with a TypeSafe-compatible endpoint, document answers with source passages, and a ViZDoom demo that picks actions from structured game state.

Jev Models & Reimplementations#community#python#local

Community

deepanwadhwa/OpenDecisionPython

SemIf-OpenJev

4.8k

An open-model semantic-if engine positioned as a local Jev-style decision stack, separate from the original SemIf repo but sharing the same author.

Jev Models & Reimplementations#community#python#open-models

Community

TheoLeeCJ/SemIf-OpenJevPython

jevos

1.3k

A local yes/no decision model that speaks Jev's wire format on a laptop CPU, refusing Choice and Score until those question types ship.

Jev Models & Reimplementations#community#python#open-models

Community

feder-cr/jevC++

Rizzo Flow

877

A local server that reads next-token probabilities from a fine-tuned Spark model and returns Choice, Score, and Noul answers with zero generated tokens.

Jev Models & Reimplementations#community#python#open-models

Community

Rizzo-AI-Academy/rizzo-flowPython

mini-Jev

A preregistered measurement of a Jev-style interface on frozen Qwen3-4B: read the option letter's logits instead of generating grammar-constrained JSON.

Jev Models & Reimplementations#community#python#local

Community

r-ms/mini-jevPython

From the Jev guides

Original write-ups on what this entry is for.

Exporting Decision Models: GGUF, ONNX, and Device Runtimes

Oct 10, 2026

Taking a decision checkpoint out of Python: what the GGUF and browser WebGPU exports preserve, what they change, and the parity evidence the model cards publish.

Jev Tools & CLIs#open-models#local#onnx

Read article

Choosing an Open Decision Model

Oct 10, 2026

Interface, size, license, claims — in that order. A selection guide across the system-one catalog, from a 706K form scorer to 27B Jev students.

Jev Practices & Patterns#open-models#evaluation#benchmarks

Read article

What Is Actually on Hugging Face

Oct 8, 2026

Hundreds of repositories carry the system-one tag, but the ecosystem is a dozen ideas in different sizes. A tour of the families, the licenses, and what a download does not include.

Jev Models & Reimplementations#open-models#hugging-face#local

Read article

Back to all resources

From the Jev community

Posts from builders shipping with Jev right now.

Follow @typesafeai