Mamba-3发布：以推理效率为核心的新型状态空间模型
原标题：Mamba-3
AI 摘要
Together AI 与卡内基梅隆大学、普林斯顿大学、Cartesia AI 合作发布了 Mamba-3，这是一种以推理效率为核心的新型状态空间模型（SSM），通过更富表现力的递推公式、复数值状态跟踪和 MIMO 变体，在 1.5B 规模下实现了比 Mamba-2、Gated DeltaNet 和 Llama-3.2-1B 更低的预填充和解码延迟。该模型还开源了基于 Triton、TileLang 和 CuTe DSL 的内核。Mamba-3 的发布反映了行业从训练优先向推理优先的转变，旨在推动质量-效率前沿。
正文节选
Mamba-3 is a new state space model (SSM) designed with inference efficiency as the primary goal — a departure from Mamba-2, which optimized for training speed. The key upgrades are a more expressive recurrence formula, complex-valued state tracking, and a MIMO (multi-input, multi-output) variant that boosts accuracy without slowing down decoding. The result: Mamba-3 SISO beats Mamba-2, Gated DeltaNet, and even Llama-3.2-1B (Transformer) on prefill+decode latency across all sequence lengths at th