Daily News · 1 min read

Hugging Face AI Updates: June 2, 2026

hugging-face jetbrains mixture-of-experts code-generation open-source llm

1. JetBrains Releases Mellum2, a 12B MoE Model for Code and Agent Workloads

Hugging Face. JetBrains published Mellum2, an open-source 12B Mixture-of-Experts model that activates only 2.5B parameters per token, delivering roughly 2x faster inference than comparably-sized dense models. The model targets high-frequency tasks within larger AI systems — routing, RAG pipelines, sub-agent planning, and private deployment on proprietary codebases — rather than serving as a general-purpose frontier model. Mellum2 is licensed under Apache 2.0 and available on Hugging Face. Source