Block releases voice model that can run locally on a MacBook M4 Max
Block announced a new voice model, according to a TechCrunch report published October 24, 2026. Block stated that the model can run locally on a MacBook M4 Max.
Why it matters
Local inference of a voice model on consumer laptop hardware lets builders run speech features without per-call cloud API costs or sending audio to third-party servers.
Who should care
Developers building voice apps, transcription or voice agent tools, and online businesses using voice features in customer support or content production.
What you can do
Benchmark the model on your own audio samples and compare its latency and output quality against your current cloud voice API before switching.