Cerebras deploys Moonshot Kimi K3.1 Pro

Moonshot released Kimi K3.1 Pro, a 2.8-trillion-parameter open-weight model featuring frontier-level coding capabilities. In an illustrative deployment, Cerebras examines how the model's Kimi Delta Attention and Gated DeepSeek Sparse Attention layers shape memory budgets and how its weights are distributed across wafers using pipeline parallelism. Open article →