LLM Architecture KB

Models

Production models

Architecture spec sheets for disclosed production LLMs. Closed or undocumented models appear with empty architecture fields per §9 citation policy.

Alibaba (Qwen Team)

Allen Institute for AI (AI2)

Cohere

Command R+ 104B 2024-04-04 · 104B · documented

DeepSeek-AI

Google DeepMind

Meta

Microsoft Research

Phi-4 14B 2024-12-12 · 14B · documented

MiniMax

Mistral AI

Moonshot AI

NVIDIA

Nemotron 3 Nano 30B-A3B 2025-12-15 · 30B (3.5B active) · documented

OpenAI

Tencent

Hunyuan-Large 389B 2024-11-05 · 389B (52B active) · open

Tsinghua KEG / Zhipu AI

GLM-130B 2022-10-05 · 130B · open

Zhipu AI

GLM-4.5 2025-07-28 · 355B (32B active) · open

Zhipu AI / Tsinghua KEG

GLM-4 9B 2024-06-04 · 9B · open