tarasovcode
Models1 min readMarkTechPost

Block Unveils Dessa, a Full-Duplex Speech Model Trained Only on Speech

Block’s AI lab released Dessa, a full-duplex speech-to-speech model trained only on human conversation audio. It runs locally on a MacBook Pro M4 Max, but Block has published no size, benchmarks or weights.

Save
Share:VKTelegramWhatsApp
Block Unveils Dessa, a Full-Duplex Speech Model Trained Only on Speech
Illustration: Tarasov Code

In brief

  • What: Block Unveils Dessa, a Full-Duplex Speech Model Trained Only on Speech
  • When: October 29, 2026 at 10:50
  • Source: MarkTechPost (media)
  • Topic: Models

#Models

Prepared by Ilya Tarasov’s Tarasov Code newsroom based on a report by MarkTechPost.

Read next

MarkTechPost · October 29, 2026NVIDIA Releases Nemotron 3.5 Transcribe: An Open 30B MoE Model for 40-Language Transcription with ~4B Active ParametersZDNet AI · October 28, 2026These Claude Enterprise features let you track AI ROI – what’s newMarkTechPost · October 28, 2026Polymathic AI and Princeton Researchers Release FlashMap, a Flow Map Drafter for 6.14x Faster LLM DecodingClaude Code releases · October 27, 2026Claude Code 2.1.312