Updates on Mixture of Experts

Every AI story we track on Mixture of Experts: 15 stories so far, each summarized in our own words and linked to the publisher that reported it.

Pulled from 125 sources

This month

Machine Learning9 min read

Mixture-of-Experts kernel engineering on Amazon SageMaker AI

AWS Blog
Machine Learning6 min read

Alibaba just released Qwen3.8-Flash: “An early preview of the architecture in Qwen4”

The New Stack (AI)
Machine Learning9 min read

Meet FreeToken: An Edge-Native MoE Serving Engine that Runs 753B GLM-5.2 on a Single Workstation GPU

MarkTechPost
Machine Learning5 min read

Cursor Open-Sources Mixture-of-Kittens (MoK): A Deterministic MoE Training Megakernel for GB300 NVL72 Racks

MarkTechPost
Machine Learning5 min read

Thinking Machines Lab Releases Inkling-Small: A 276B Total, 12B Active Open Weights Multimodal MoE Model

MarkTechPost

AMD Releases Instella-MoE-16B-A3B: A Fully Open Mixture-of-Experts LLM With 2.8B Active Parameters Trained On Instinct GPUs

MarkTechPost
Machine Learning4 min read

DeepSeek Upgrades DeepSeek-V4-Flash-0731 with Major Agentic and Coding Gains

MarkTechPost
Machine Learning5 min read

Moonshot AI Open-Sources MoonEP: A Perfectly Balanced Expert Parallelism Library for MoE Training

MarkTechPost
Machine Learning5 min read

Moonshot Opens Kimi K3 Weights Under a Revenue-Tiered License

Unite.AI

Introducing Mellum2: A 12B Mixture-of-Experts Model by JetBrains

Hugging Face

EMO: Pretraining mixture of experts for emergent modularity

Hugging Face
That's everything we have on Mixture of Experts right now