# Perplexity — Optimizing On-Device Inference for Apple Silicon

- Company: Perplexity (perplexity.ai)
- Announced: 2026-09-01
- Category: developer-tool-release
- Coverage: not counted
- Announcement: yes
- Group: announcements
- Source: https://www.perplexity.ai/hub/blog/optimizing-on-device-inference-for-apple-silicon
- Record: https://forck.live/items/10757-optimizing-on-device-inference-for-apple-silicon
- Subject: Perplexity
- Models affected: Qwen3.6-35B-A3B

Perplexity released Lily, a lightweight local inference engine built specifically for Apple silicon and the Qwen3.6-35B-A3B model, with separate optimizations for prefill and decode phases. Benchmarked on an M5 Max MacBook Pro, Lily achieves 1.23× prefill throughput and 1.35× decode throughput compared to MLX-LM, reaching 5,749.9 prefill tokens per second and 186.6 decode tokens per second at 4K-token prompt and context lengths. A standalone demo is publicly available on GitHub.

## Evidence

Verbatim from https://www.perplexity.ai/hub/blog/optimizing-on-device-inference-for-apple-silicon:

> Lily, our lightweight local inference engine, is built specifically for Apple silicon and Qwen3.6-35B-A3B, with separate optimizations for prefill and decode. A standalone demo is publicly available on GitHub.

---

Record: https://forck.live/items/10757-optimizing-on-device-inference-for-apple-silicon
Catalogue: https://forck.live/llms.txt
Current issue: https://forck.live/feed.md