Reproduce “Apriel Voice 27B Preview: An Open Model for Speech, Reasoning, and Tool Use” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Aman Tiwari, Akshay Kalkunte, Akintunde Oladipo, Jash Shah, Tara Bogavelli, Hoang Nguyen, Nishanth Madhusudhan, Vishal Juneja and 23 more
Apriel Voice routes speech directly into the Gemma 4 26B-A4B reasoning model, alongside text and images, to generate responses or tool calls; application software executes tools and may synthesize response text. In τ-Voice thinking mode, it reports 65.51% task success and 1.91 seconds latency, versus 67.78% and 3.01 seconds for Parakeet + Gemma 4 26B: lower latency, not higher success.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.