#Mixture-of-Experts

#1

GLM‑4.5

Open-source hybrid‑reasoning MoE foundation model optimized for intelligent agent tasks with 128K context and tool use.

#2

DeepSeek V3

Open-source mixture-of-experts model offering GPT-4o-level reasoning at a fraction of the cost.

#3

MiniMax‑M1

Open‑source large‑scale reasoning model with 1 million token context and hybrid Mixture‑of‑Experts architecture.