{"id":"higgs-realtime-duplex","slug":"higgs-realtime-duplex","name":"Higgs Realtime Duplex","orgId":"bosonai","family":"higgs-realtime-duplex","releaseDate":null,"availability":"announced","openWeights":false,"license":null,"paramsTotal":null,"paramsSource":null,"contextWindow":null,"knowledgeCutoff":null,"modelCardUrl":"https://www.boson.ai/blog/higgs-realtime-duplex","hfRepo":null,"modalities":["audio","video","text"],"aliases":["Higgs Realtime Duplex","higgs-realtime-duplex","higgs realtime duplex","boson higgs realtime duplex","higgs duplex"],"aliasNote":null,"origin":"tracked","tracked":{"trackedSince":"2026-11-05","announced":"2026-11-05","blurb":"Boson AI's research preview of the next generation of Higgs Realtime: a full-duplex interactive omni model that listens, watches and talks at the same time while rendering a live face, and that can think, call tools and delegate to a larger backend model without pausing the conversation. Front model is a 45B MoE trained on about 2 trillion tokens of text, audio and video, running in 160 ms chunks. An API for Duplex is on a waitlist; the current Higgs Realtime speech-to-speech API remains available separately.","api":"waitlist","openWeights":"no","params":{"text":"45B MoE (front model)","source":"Boson AI blog"},"facts":[{"label":"Role","value":"Full-duplex interactive omni model: continuous audio and video in, speech and a rendered face out, with background thinking and tool use"},{"label":"Chunking","value":"Streams cut into 160 ms chunks for input (voice, camera) and output (voice, face, background work)"},{"label":"Architecture","value":"45B mixture-of-experts front model; can hand hard turns to a larger backend model with shared context while keeping the conversation going"},{"label":"Training","value":"Continued training on about 2 trillion tokens of text, audio and visual data; multi-teacher on-policy distillation and a recursive self-improvement loop (per Boson)"},{"label":"Face and voice","value":"Zero-shot: a short reference audio clip sets the voice and a reference image sets the face at deployment time (not baked into the weights)"},{"label":"Access","value":"Research preview / demo; Duplex API on waitlist (\"coming soon\" per the Nov 5 blog). Current Higgs Realtime remains the hosted speech-to-speech API"},{"label":"Weights","value":"None stated for Duplex"}],"claims":["Lab claim: Big Bench Audio accuracy 98.9; Audio MultiChallenge APR 44.9; Full-Duplex-Bench v1.5 average 75.5 (Boson's published comparison table).","Lab claim: QIVD streaming accuracy 56.0; ProactiveVideoQA PAUC@ω=0.5 of 31.9; τ-Voice task completion pass@1 69.8%; IFEval accuracy 94.1.","Lab claim: design target is that an hour with the model costs at most one-tenth as much as an hour of a person's time."],"sources":[{"kind":"lab","title":"Higgs Realtime Duplex blog (Boson AI)","url":"https://www.boson.ai/blog/higgs-realtime-duplex","date":"2026-11-05"},{"kind":"lab","title":"Boson AI on X: Higgs Realtime Duplex","url":"https://x.com/boson_ai/status/2118403814403580173","date":"2026-11-05"}],"firstBenchmark":null},"organization":{"id":"bosonai","name":"Boson AI","slug":"bosonai","country":null,"url":"https://www.boson.ai","color":"#A39E8F","mark":"BO"},"price":null,"answer":{"word":"Pending benchmarks.","scope":"Announced Nov 5, 2026; no independent results yet.","lead":"As of Nov 22, 2026, Higgs Realtime Duplex is pending benchmarks on isitsota: announced Nov 5, 2026, with no independent results on the boards we track yet.","confidence":"low"},"status":"PENDING_BENCHMARKS","verdicts":[]}