Perplexity 已开源 Lily，这是一个专为 Perplexity Computer 产品中的混合计算设计的本地推理引擎。Lily 专门针对在 Apple silicon 上运行 Qwen3.6-35B-A3B 模型进行了优化，将其视为一个独立的平台。基准测试显示，在 M5 Max MacBook Pro 上，Lily 在 prefill 和 decode 吞吐量方面均优于 MLX-LM，在保持输出质量的同时，有效地处理了 prefill 和 decode 工作负载的不同计算需求。 AI
影响 能够更高效地在 Apple Silicon 上进行本地 LLM 推理，可能提高 Perplexity Computer 等应用程序的混合计算性能。
排序理由 为本地硬件开源专用推理引擎。
AI 生成摘要 · Google Gemini · 来自 5 个来源。 我们如何撰写摘要 →
完整方法见我们的编辑标准。
Lily treats Apple silicon as a distinct inference platform and maps Qwen’s operations directly to its compute and memory architecture. Lily is now available: https://t.co/WbawYlTqSe
We benchmark Qwen3.6-35B-A3B on an M5 Max MacBook Pro. Across ten prompt lengths and ten decode contexts, Lily was faster than MLX-LM, averaging 1.23× higher prefill throughput and 1.35× higher decode throughput while keeping output quality effectively unchanged. https://t.co/K7…
Lily treats prefill and decode as fundamentally different workloads. Prefill processes many prompt tokens at once and can reuse weights across them, while decode generates one token at a time with much less reuse, making memory traffic and bandwidth more important. https://t.co/…
Hybrid compute splits work between cloud models and a local model on the Mac. Local inference must keep pace with the rest of the task. MLX-LM is a general-purpose framework, while Lily is purpose-built for this inference workload. https://t.co/MFsA8LgK0O
Today we’re open-sourcing Lily, the local inference engine we built for hybrid compute in Perplexity Computer. Lily is specialized for Qwen3.6-35B-A3B on Apple silicon, built so on-device compute doesn’t bottleneck Computer tasks. Read more: https://t.co/OnowTP3ql6