Motif 3 opens a 314B MIT model

Motif Technologies has released Motif 3, a 314-billion-parameter Mixture-of-Experts model with 13.2 billion active parameters per token, a 256K context window, and an MIT license. The Motif 3 model card says the open-weight model was trained on roughly 12.5 trillion tokens and is tuned for long-horizon tool use; its technical report documents the architecture and evaluation suite. Teams comparing Meta’s 30B local agent weights should notice the split: Motif’s MIT terms invite commercial modification, but the official serving recipe sets data parallelism to eight on H200 or B200 GPUs, so “open” still carries a serious deployment bill.