AMD Releases Fully Open-Source MoE Model Instella

AMD launched Instella-MoE, its first fully open-source Mixture-of-Experts language model with 16 billion total parameters and 2.8 billion active parameters per token, alongside a fully disclosed training pipeline.

2026-07-28 ~ 2026-07-30 · 2 related posts

• AMD opens Instella-MoE, a fully open 16B MoE model trained on MI300X and MI325X — QuixiAI · 2026-07-28

1 near-duplicate retellings: BanghuaZ