Published source analysis, not our speech-model execution
Two-track examples, timing and actual access →
The release-conversation example has two speaker tracks and post-added captions.
Block
Dessa
Block’s full-duplex speech research demonstration claims audio-only training and local inference on MacBook Pro M4 Max or better. Download and waitlist remain prospective; the page does not specify public weights, model licensing or API terms, and the three boards have no comparable result.
Original model explanation ↗No comparable leaderboard score is included yet. Explore the evaluations and examples below.
Evaluations: conditions and results
Publisher-selected recordings and post-added captions, not our audio evaluation
Official two-track samples and current research account ↗
How this capability connects to works and uses
Research showing, not customer-use or business-outcome evidence
Original conversation examples ↗
Block model evolution
- Dessa · 全双工语音研究展示
The original X announcement is October 23 at 18:23 UTC, October 24 in Beijing. The page presents recordings and an M4 Max claim; general downloads/waitlist remain prospective, without establishing open weights.
Original release ↗