Tagged with

4 articles found

LG’s K-EXAONE: Korea’s 236B MoE Gamble Challenges Global AI Hierarchy

LG AI Research’s K-EXAONE claims 104% performance vs Qwen3 with only 23B active parameters, but the real story is in the architectural tradeoffs and Korea’s AI sovereignty push

#benchmark-wars#korean-tech#mixture-of-experts

commercial-license

Korea’s AI Moonshot: How Solar-Open-100B Declares War on U.S. Model Hegemony

South Korea’s Upstage just dropped a 102-billion-parameter model with a commercial license, trained on government GPUs, and aimed squarely at breaking America’s AI stranglehold.

#commercial-license#korean-tech#mixture-of-experts...

large-language-models

The Open-Source Tipping Point: INTELLECT-3 Proves 100B+ MoE Models Can Outperform Corporate Giants

Prime Intellect releases a 100B+ parameter Mixture-of-Experts model that beats larger frontier models in reasoning, math, and coding, and they’re giving away the entire training recipe.

#large-language-models#mixture-of-experts#open-source...

coding-agents

MiniMax M2: The Open-Source Coding Agent That’s Actually Affordable

MiniMax’s 229B-parameter MoE model delivers Claude-level coding performance at 8% the price, challenging the economics of agentic development.

#coding-agents#llm-efficiency#mixture-of-experts...