AMD releases Instella-MoE, a 16B mixture-of-experts language model, under a research-only license

AMD introduced Instella-MoE in a ROCm blog post dated July 24, 2026: a mixture-of-experts language model with 16 billion total and 2.8 billion active parameters, trained from scratch on AMD Instinct MI300X and MI325X GPUs with the Primus and Miles frameworks. AMD says it is releasing model weights from every training stage, training configurations, data mixtures and code, and the Hugging Face checkpoints run from Instella-MoE-16B-A3B-Pretrain to the reinforcement-learning-tuned Instella-MoE-16B-A3B-Think. AMD calls the model fully open, but the model card says the models are licensed for academic and research purposes under a ResearchRAIL license and are released for research purposes only. USASI has not tested the models.12

USASI. (2026, October 13). AMD releases Instella-MoE, a 16B mixture-of-experts language model, under a research-only license. United States of America Superintelligence. https://unitedstatesofamericasuperintelligence.com/news/amd-releases-instella-moe/

@misc{usasi_news_amd_releases_instella_moe,
  author = {{USASI}},
  title = {AMD releases Instella-MoE, a 16B mixture-of-experts language model, under a research-only license},
  year = {2026},
  month = oct,
  howpublished = {\url{https://unitedstatesofamericasuperintelligence.com/news/amd-releases-instella-moe/}},
  note = {United States of America Superintelligence. Last updated 2026-10-13}
}