project-maya
Project Maya - run GLM-5.3-Flash (321B MoE) privately on your own GPU(s), from one card to 16, NVIDIA or…
project-maya launched on 6 October 2026. It ranked #98 on Daily Launch on 30 October 2026 with a Launch Score of 22, from +70 stars today on GitHub. It was new in the ranking that day.
What is Project Maya?
Models this size usually need a datacenter. Maya runs GLM-5.3-Flash (zai-org, MIT license; 321 B parameters, about 18 B active per token, a context of up to 1 M tokens) and DeepSeek-V4.1-Flash-1017 (DeepSeek, MIT license; 748 B parameters) on the hardware you already have. Each model has an engine built for it. It keeps the experts your conversation uses most on the GPU, the next ones in RAM and the rest on your NVMe SSD, and moves them as you chat.
From its README on GitHub
Project Maya (project-maya) is a repo by mw00 on GitHub. It is written in Cuda and released under the MIT license.
How to get Project Maya
Get it from its repository on GitHub.
In its own words
Project Maya - run GLM-5.3-Flash (321B MoE), DeepSeek-V4.1-Flash-1017 (748B MoE) and MiniMax-M3.1-Flash (428B MoE) privately on your own GPU(s), from one card to 16, NVIDIA or AMD, Linux or Windows. Engine, server, dashboard, OpenAI- and Anthropic-compatible API, quant tools.
From its GitHub repository