audio.cpp 0.4 Adds Voice Models and Boosts VRAM Efficiency

audio.cpp version 0.4 adds support for five new voice models, including Higgs Audio v3 TTS, and elevates GGUF to a first-class format across 35 model families. This update also delivers significant performance improvements, reducing VRAM usage by up to 37%.

2026-07-24 ~ 2026-07-24 · 2 related posts

• audio.cpp 0.4 adds 5 speech models and reports up to 37% VRAM savings — Acceptable-Cycle4645 · 2026-07-24

1 near-duplicate retellings: Acceptable-Cycle4645