70 of 70
| Full-duplex | Weights | Commercial | Languages | |||
|---|---|---|---|---|---|---|
 Amazon Nova 2.5 Sonic APIAmazon / AWS | Full-duplex | 2026-10 | — | — | Yes | — |
| Full-duplex | 2026-10 | — | — | Yes | en es fr +17 | |
 Eleven Realtime (v1 Mini / v1 Max) APIElevenLabs | Cascade | 2026-10 | — | — | Yes | 90+ |
Phonic-1Phonic | Full-duplex | 2026-10 | — | — | Yes | 51 |
Rime PrismRime | Half-duplex | 2026-10 | — | — | Yes | en |
Mistral Voice Agents (Voxtral 3)Mistral AI | Cascade | 2026-10 | — | — | Yes | — |
 Cartesia Managed AgentsCartesia | Cascade | 2026-10 | — | — | Yes | — |
LFM3-Audio-2B / LFM3-Omni-2BLiquid AI | Half-duplex | 2026-10 | — | open | Unclear | en +17 |
Apriel Voice 27B PreviewServiceNow | Half-duplex | 2026-10 | — | open | Yes | en + |
| Eleven Realtime TranslationElevenLabs | Other | 2026-10 | — | — | Yes | 90+ |
 Griffin-LiteTavus | Full-duplex | 2026-10 | — | — | Unclear | — |
Realtime-VenusAnt Group (inclusionAI) | Full-duplex | 2026-09 | — | open | Yes | en zh |
GanderOmni-Interaction-Gander | Full-duplex | 2026-09 | — | open | Yes | — |
 Gemini 3.8 Live APIGoogle DeepMind | Full-duplex | 2026-09 | — | — | Yes | en ja zh +94 |
 StepAudio 3 Realtime APIStepFun | Full-duplex | 2026-09 | — | — | Unclear | — |
 Qwen-Omni Realtime (Qwen3.8-Omni-Flash-Realtime)Alibaba Cloud | Full-duplex | 2026-09 | — | — | Yes | 113 |
| Gemini 3.8 Flash TTS / Flash-Lite TTSGoogle DeepMind | Other | 2026-09 | — | — | Yes | 100+ |
| Qwen3.8-LiveTranslateAlibaba Cloud | Full-duplex | 2026-09 | — | — | Yes | 60 |
| Gemini 3.8 Live with Live AvatarGoogle DeepMind | Full-duplex | 2026-09 | — | — | Yes | 97 |
| Cascade | 2026-09 | — | — | Yes | 42+ | |
| Full-duplex | 2026-08 | 448 ms | open | Yes | en | |
 SeedRealtimeByteDance Seed | Full-duplex | 2026-08 | — | — | Unclear | — |
| Qwen-Audio Realtime (3.0 / 3.1) APIAlibaba Cloud | Full-duplex | 2026-08 | — | — | Yes | zh en ja +11 |
 Higgs RealtimeBoson AI | Full-duplex | 2026-08 | 700 ms | — | Yes | 100+ |
| Cartesia Sonic-3.6Cartesia | Other | 2026-08 | 90 ms | — | Yes | en ja zh +41 |
DuplexOmniXi'an Jiaotong University / Peking University / Meituan | Full-duplex | 2026-07 | 506 ms | open | Yes | zh en |
Lychee-FDHITsz-TMG | Full-duplex | 2026-07 | — | open | Yes | zh en |
| Full-duplex | 2026-07 | — | — | Yes | en ja zh +52 | |
 GPT-Live 1 APIOpenAI | Full-duplex | 2026-07 | 798 ms | — | Yes | — |
 PolyAI Dialog-RSN-1PolyAI | Cascade | 2026-07 | 300 ms | — | Yes | en |
| Qwen Audio AgentQwenAudio | Cascade | 2026-07 | — | — | Yes | — |
BayLing-DuplexICT, Chinese Academy of Sciences | Full-duplex | 2026-06 | — | open | Yes | — |
 TML-Interaction-SmallThinking Machines Lab | Full-duplex | 2026-05 | 400 ms | — | Unclear | en |
| GPT-Realtime-TranslateOpenAI | Full-duplex | 2026-05 | — | — | Yes | en ja zh +67 |
 AVTR-1Avaturn | Cascade | 2026-05 | — | open | Unclear | — |
 Raon-SpeechChat-9BKRAFTON | Full-duplex | 2026-04 | — | open | Non-commercial | en |
 AssemblyAI Voice Agent APIAssemblyAI | Cascade | 2026-04 | — | — | Yes | — |
| Full-duplex | 2026-03 | — | open | Research only | zh en | |
| Gemini 3.1 Flash Live PreviewGoogle DeepMind | Full-duplex | 2026-03 | — | — | Yes | en ja zh +21 |
 Runway CharactersRunway | Cascade | 2026-03 | 1750 ms | — | Yes | — |
 MiniCPM-o 4.5OpenBMB / ModelBest | Full-duplex | 2026-02 | — | open | Yes | en zh +30 |
MOSS-TTSOpenMOSS (Fudan) | Other | 2026-02 | — | open | Yes | zh en de +17 |
 PersonaPlexNVIDIA (ADLR) | Full-duplex | 2026-01 | 170 ms | open | Yes | en |
| Qwen3-TTSAlibaba Cloud | Other | 2026-01 | — | open | Yes | en zh ja +7 |
Avatar ForcingKAIST / NTU Singapore / DeepAuto.ai | Cascade | 2026-01 | — | open | Non-commercial | — |
| Gemini 2.5 LiveGoogle DeepMind | Full-duplex | 2025-12 | — | — | Yes | en ja zh +21 |
| Amazon Nova 2 SonicAmazon / AWS | Full-duplex | 2025-12 | — | — | Yes | en es fr +4 |
| SALMONN-omniByteDance · Tsinghua · Cambridge | Full-duplex | 2025-11 | — | — | Unclear | en |
| Cascade | 2025-11 | — | — | Yes | — | |
| Qwen3-OmniAlibaba Cloud | Full-duplex | 2025-09 | — | open | Yes | en zh ja +16 |
| Step-Audio 2 miniStepFun | Full-duplex | 2025-08 | — | open | Yes | zh en |
 Kyutai UnmuteKyutai | Cascade | 2025-07 | — | open | Yes | en fr |
| ElevenLabs Agents (ElevenAgents)ElevenLabs | Cascade | 2025-05 | — | — | Yes | 90+ |
 Kimi-AudioMoonshot AI | Full-duplex | 2025-04 | — | open | Yes | zh en |
LLaMA-Omni 2ICT / CAS (ictnlp) | Half-duplex | 2025-04 | 583 ms | open | Research only | en |
 Sesame CSMSesame AI | Full-duplex | 2025-03 | — | open | Yes | en |
| Qwen2.5-OmniAlibaba Cloud | Half-duplex | 2025-03 | — | open | Yes | en zh ja +1 |
| HibikiKyutai | Full-duplex | 2025-02 | — | open | Yes | fr en |
 Deepgram Voice AgentDeepgram | Cascade | 2024-12 | 425 ms | — | Yes | 60+ |
| CosyVoice 2Alibaba Tongyi | Other | 2024-12 | — | open | Yes | zh en ja +1 |
 GLM-4-VoiceZhipu AI (Z.ai) | Full-duplex | 2024-10 | — | open | Unclear | zh en |
| MoshiKyutai | Full-duplex | 2024-09 | 160 ms | open | Yes | en |
 LiveKit AgentsLiveKit | Cascade | 2024-09 | — | — | Yes | en |
 HF speech-to-speechHugging Face | Cascade | 2024-08 | — | open | Yes | en fr es +3 |
 UltravoxUltravox (Inworld) | Half-duplex | 2024-08 | — | open | Yes | en +41 |
 Retell AIRetell | Cascade | 2024-02 | 600 ms | — | Yes | 55+ |
 PipecatDaily | Cascade | 2023-12 | — | — | Yes | en |
 BlandBland AI | Cascade | 2023-11 | 200 ms | — | Yes | 40+ |
 SeamlessM4T v2Meta AI | Other | 2023-11 | — | open | Non-commercial | 100+ |
 VapiVapi | Cascade | 2023-07 | 500 ms | — | Yes | en ja es +100 |
