Bland unveils Tofu-1, a full-duplex speech-to-speech model built for noisy high-stakes calls
JaynitMakwana · x · 2026-11-14
Bland introduced Tofu-1, a full-duplex speech-to-speech model designed for the worst call conditions: moving cars, hospital rooms, weak signal, and multiple people talking at once.
- Listens while speaking: callers can interrupt or correct mid-sentence
- Reasons while talking: keeps the conversation flowing while looking things up, and never claims a task is done before it is
- Speaker separation: distinguishes the caller from noise like a TV or crying baby
- Generates whole scenes: produces multiple speakers, overlapping voices, and ambient room sound
Demos include a car call with background chatter and a support agent talking while retrieving info. Research preview access is open to a limited number of enterprises.
Related event: Bland launches full-duplex speech model Tofu-1 for noisy, high-stakes calls(2 posts)→
More from Multimodal
- Seedance 2.5 demo: a 1080P split-scene clip with prompts shared — LudovicCreator · 2026-11-14
- Matthew Berman open-sources Scarlet, an AI video editing agent that speeds up his workflow — MatthewBerman · 2026-11-14
- AI-generated 'red thread' visual narrative showcases cinematic consistency — tallmetommy · 2026-11-14
- Bland unveils Tofu-1, a full-duplex speech-to-speech model built for noisy high-stakes calls — Scobleizer · 2026-11-14
- 2orbs TV, a 24/7 AI-made music video channel, launches on Apple TV — azed_ai · 2026-11-14
- Open 8B model, 61 style templates: 10 photo restyles with no LoRA — Single-Climate2259 · 2026-11-14