NAYHEIN

AI

奈亨智能

Pursuing Mercury-level efficiency in AI reasoning — achieving autoregressive reasoning with diffusion-speed generation.

We're building an open-source implementation of breakthrough AI architectures that combine the best of autoregressive and diffusion models.

Our Mission

Why We're Building This

Mercury Coder 2 demonstrated that autoregressive models can achieve diffusion-level speeds while maintaining coherent reasoning. We believe this architecture should be accessible to everyone.

10x

Faster inference than traditional autoregressive models

Open

Research and implementation for the community

Replicable

Documented approach for others to follow

Technical Approach

How We're Getting There

1

Architecture Analysis

Studying Mercury's parallel decoding mechanism and understanding how it achieves diffusion-style generation while preserving autoregressive reasoning capabilities.

2

Implementation

Building our version from first principles, focusing on the key innovations: parallel token prediction, efficient attention patterns, and coherent multi-step reasoning.

3

Open Release

Sharing our findings, code, and models with the community so others can build upon efficient reasoning architectures.