
Innovation and high quality speaks for itself, @boson_ai #higgsaudio v2 is a big hit! treading on top 5 on @huggingface this week.

CPO, Boson.AI

Innovation and high quality speaks for itself, @boson_ai #higgsaudio v2 is a big hit! treading on top 5 on @huggingface this week.

The hardest part of a conversation isn't talking. It's knowing when not to. In this demo, Higgs Realtime Duplex is asked to check the markets and a training run. It stays quiet while the driver orders at the drive-thru, reports back once the order is done, and at the pickup window confirms the driver never ordered animal-style fries. It heard every word and knew which ones were meant for it. How it works: ➡️ It decides for itself when to speak, and recognizes when a remark isn't addressed to it. ➡️ Tool calls run in the background, so it can check information while the conversation carries on. Read more about Higgs Realtime Duplex (research preview): boson.ai/blog/higgs-realtime… #ai #voiceai #aiagents #conversationalai #bosonai

What if an AI could join your team meeting and actually keep up? In this demo, Higgs Realtime Duplex presents an architecture, takes questions mid-sentence, reads the whiteboard and updates a diagram as the conversation moves. How it works: instead of waiting for its "turn," the model processes everything in 160-millisecond chunks. Your voice and camera stream in while its voice and face stream out, all at the same time. That's what lets it handle interruptions naturally, the way a person would. Higgs Realtime Duplex is now available as a research preview. Read more: boson.ai/blog/higgs-realtime… #ai #voiceai #realtimeai #conversationalai #bosonai

The AI agent can talk, listen, think and do work at the same time, join the engineering team meeting for real. youtu.be/vclcTznkNjM

Introducing Higgs Realtime Duplex, a research preview of the next generation of Higgs Realtime and Boson AI's first full-duplex interactive omni model. It listens, watches, and talks at the same time, and it has a face of its own. #ai #nextgen #bosonai #fullduplex #voiceagent

We'll be spending a lot more time trying to understand the outputs of language models. A few thoughts, tips & tricks: Writing. Something I've had success with: Ask your LLM to explain something in ASD-STE100, it's a controlled language specification originally developed for aerospace maintenance documentation. LLMs well-versed in this language and it comes with heavy constraints on clean writing style that I often find a lot more readable. Sometimes I've tried to soften it a bit e.g. ask for "80% of the way to ASD-STE100" because the spec is quite stringent. But even better: Diagrams / images. Instead of writing, ask your LLM to create a diagram. These can be a lot easier to process, parse, and understand. But even better: Web pages. Ask for output "in HTML" to get a beautiful, interactive webpage. LLMs are getting really good at frontend and can create beautiful experiences, animations, etc. But even better: Explainer videos. The output format I am most bullish on is fully custom / bespoke explainer videos generated on any arbitrary topic. Experiment with things like "Create a 3b1b style video explainer on X. Use my ElevenLabs API key for audio narration". (you'd need an API key for the latter or you can ask your LLM to find you decent free alternatives that use your local compute). This is actually starting to work! In summary: - As LLMs get better, they will do more and more of the legwork autonomously, and a lot more of our work will rise up the abstractions into oversight and understanding. - Luckily, LLMs can help here too because as intelligence and code are increasingly abundant, you can ask for large, custom, discardable software artifacts (e.g. web apps, video explainers) that would have never made sense to create before. Push the boundaries here and you'll be surprised.

Real-time voice AI is easy to demo but much harder to build. We put together a hands-on tutorial for Higgs Realtime that walks through the pieces needed to build one by yourself. boson.ai/blog/higgs-realtime…

AT&T is doing exactly what OpenAI and Anthropic fear about enterprise. they’re routing 40% of employee AI usage to open models, and intend to boost it to 60-70%. > coding costs down 56% > quality dropped just 2% > 45B tokens/day btw. AT&T AI chief: open models are “just as good or better” for many tasks. they still use frontier models for the critical work but everything else gets routed to cheaper open models.

#higgs #realtime brings frontier-level voice intelligence into production—fast, natural, multilingual, and cost-efficient enough to scale. A major step toward AI that doesn’t just respond, but listens, adapts, and gets work done.
Meet Higgs Realtime: a cost-efficient, real-time speech-to-speech model. It listens while it talks. Interrupt it mid-sentence, change your mind, and it keeps up, all the way through real workflows. Available today 🧵

we pushed the code to production, @boson_ai Avatar Generation API is live now Create a video - docs.boson.ai/api-reference/…

Check out this replay of our office hour with @boson_ai, where you can see how Higgs Audio v3 TTS is served on SGLang-Omni! For more information, check out the deck made by @huapeng_zhou: popsoda2002.github.io/higgs_…
Check out the @sgl_project office hours about @boson_ai Higgs Audio v3 TTS, courtesy of Ke Bai and Huapeng Zhou. Worth watching the full video. youtube.com/watch?v=iIfr4oqA…

📣 SGLang Office Hour: How to Build Fast, Modern Voice Cloning with @boson_ai Our Office Hour is back! This week we're joined by Ke and Huapeng from the Boson AI team to talk about Higgs Audio V3 TTS, how to build fast, modern voice cloning that actually sounds human, and their work optimizing SGLang-Omni for TTS production serving. Bring your questions for Ke and Huapeng, see you tomorrow!

🎙️ New blog: Higgs Audio v3 TTS on SGLang-Omni: Real-Time, Controllable Speech for Voice Agents Higgs (@bosonai) is a ~4B chat-native TTS model (Qwen3-4B backbone, interleaved text + audio tokens) built for streaming voice: 🎙️ Synthesis starts before a full sentence arrives 🌍 100 languages at single-digit WER/CER 🗣️ Zero-shot voice cloning from a short reference clip, across languages 🎛️ 20+ inline control tokens for emotion, style, prosody & sound effects ⚡ Serving with SGLang-Omni Runs as a multi-stage pipeline: preprocessing → audio encoder → TTS engine → vocoder ✅ CUDA Graph decode ✅ Async one-step-lookahead ✅ Batched vocoder & encoder ✅ RadixAttention partitioned by reference audio to amortize repeated cloning 📊 Performance: 14.74 req/s at RTF 0.262 with 16 concurrent requests on a single H100 (bf16)

🎉 Meet Higgs Audio v3 TTS from @boson_ai, a ~4B chat-native TTS model for real-time voice agents. Day-0 support is live in SGLang-Omni! > Low-latency streaming > 100 languages, single-digit WER/CER > Zero-shot voice cloning from a short clip > 20+ inline tokens for emotion, style & SFX > 14.74 req/s @ RTF 0.262 on 1× H100 👉 Cookbook: sgl-project.github.io/sglang… Run it now with SGLang-Omni!
Higgs Audio v3 TTS is here. Built for voice AI that speaks, not just reads: • 100 languages with single-digit WER/CER • inline control over emotion, style, prosody, and sound effects • API, Workspace, and open weights • Blog 👉 boson.ai/blog/higgs-audio-v3… Watch the demo 👇

Higgs Audio v3 TTS is here. Built for voice AI that speaks, not just reads: • 100 languages with single-digit WER/CER • inline control over emotion, style, prosody, and sound effects • API, Workspace, and open weights • Blog 👉 boson.ai/blog/higgs-audio-v3… Watch the demo 👇

Human communication is not voice alone. We rely on expression, timing, attention, and presence to build trust and understanding. Adding a real-time human avatar brings a critical visual layer, making human-AI interactions more natural, more engaging. youtube.com/watch?v=xOnMYhca…