YC-backed open-source inference engine Magnitude tops Hacker News with up to 2x faster decoding

YC S25 startup Magnitude released an open-source inference engine that topped Hacker News. It compiles and tunes kernels on-device, with official benchmarks showing up to 2x faster decoding than llama.cpp, including a 92% gain on Metal.

2026-10-01 ~ 2026-10-01 · 2 related posts

• Magnitude (YC S25) Open-Sources Inference Engine with Up to 2x Faster Decode Than llama.cpp — petrusenko_max · 2026-10-01
• Magnitude inference engine hits #1 on HN, claims up to 2x faster local open-model runs than llama.cpp — nickbaumann_ · 2026-10-01