audio.cpp update cuts Higgs Audio memory 48%, supports 110+ models

audio.cpp shipped runtime-level optimizations: Higgs Audio TTS peak memory dropped 48% to about 6GB with unchanged speed, several TTS models got faster, and the project now supports 110+ audio models.

2026-10-08 ~ 2026-10-08 · 2 related posts

• audio.cpp cuts Higgs Audio TTS VRAM by 48%, now supports 110+ audio model families — Acceptable-Cycle4645 · 2026-10-08

1 near-duplicate retellings: Acceptable-Cycle4645