MiMo-V2.5-Pro
Flagship MoE model for long-context agents and complex coding
...Architecturally, it uses a hybrid attention system combining Sliding Window Attention and Global Attention to significantly reduce memory usage while preserving long-context performance. It also integrates multi-token prediction modules that accelerate inference and improve reinforcement learning efficiency. Trained on around 27 trillion tokens with FP8 mixed precision and refined through supervised fine-tuning, large-scale agentic reinforcement learning, and distillation.