Text to speech

Models 107

• Gen 4 Qwen
  Qwen Audio 3.0 TTS

  By Alibaba

  Qwen Audio 3.0 TTS is a production oriented text to speech model built on a 12.5 Hz speech tokenizer and a five stage progressive training pipeline. It supports zero shot voice cloning, 16 languages, 20 Chinese dialect regions, and one pass synthesis up to 3 minutes, with strong robustness to noisy or degraded reference audio.

  🔊Text to speech🗣️Voice cloning🎙️Voiceovers🎧Audiobooks
  NewAudio

  Released 7d ago
• Gen 4 GPT
  GPT Live 1 mini

  By OpenAI

  GPT-Live-1 mini is a smaller full-duplex voice model that listens and speaks simultaneously for natural, real-time conversation. It delegates web search or agentic tasks to a frontier model in the background and powers ChatGPT Voice for Free users.

  🎙Voice chatting🔊Text to speech🤖Task automation🎤Voice assistants🔎Search🌍Translations
  NewAudio

  Released 19d ago
• Gen 4 GPT
  GPT Live 1

  By OpenAI

  GPT-Live-1 is a full-duplex voice model that listens and speaks simultaneously for natural, real-time conversation. It delegates web search, reasoning, or agentic work to a frontier model in the background while keeping the conversation flowing, and powers ChatGPT Voice for Go, Plus, and Pro users.

  🎙Voice chatting🔊Text to speech🤖Task automation🎤Voice assistants🔎Search🌍Translations
  NewAudio

  Released 19d ago
• Gen 4
  Sume Avatar 1.0

  By Sume

  Sume Avatar 1.0 is a multi-agent orchestration system exposed as a single avatar video model. It routes a script, avatar choice, and optional product image across specialized image, audio, and video models to produce consistent talking head videos up to 60 seconds long.

  🎥Videos🔊Text to speech🎥Product videos👤Avatars📹UGC videos
  NewVideo

  Released 26d ago
• Gen 4 Fish
  S2.1 Pro

  By Fish Audio

  S2.1 Pro is a neural text to speech model that converts text into speech across 83 languages with automatic language detection. It supports inline bracket tags for emotion and paralinguistic control, multi-speaker dialogue, and instant voice cloning from a reference audio sample. It streams audio in real time with about 70 to 90 ms time to first audio and outputs mp3, wav, pcm, or opus.

  🔊Text to speech🗣️Voice cloning🎙️Voiceovers🎧Audiobooks
  NewAudio

  Released 1mo ago
• Gen 4 Seed
  Seed Audio 1.0

  By ByteDance

  Seed Audio 1.0 is ByteDance's universal audio generation model that creates voice, music, sound effects, and ambient soundscapes from text prompts. It supports zero-shot voice cloning from short audio references, multi-character dialogue generation in a single pass, and cross-lingual synthesis without fine-tuning. Accessible via Volcano Engine API.

  🔊Advanced audio generation🔊Text to speech🗣️Voice cloning🎶Music generation
  NewAudio

  Released 1mo ago
• Gen 3
  ZONOS2

  By Zyphra AI

  ZONOS2 is Zyphra’s open-source real-time text-to-speech model with MoE architecture, high-fidelity zero-shot voice cloning, and multilingual expressive speech generation.

  🔊Text to speech🗣️Voice cloning🎙️Voiceovers
  NewMultimodal

  Released 1mo ago
• Gen 4 Gemini
  Gemini 3.5 Live Translate

  By Google DeepMind

  Gemini Audio is Google DeepMind’s closed-source native audio model family for low-latency live dialogue, controllable speech generation, audio understanding, and voice-first applications.

  🎧Audio translation🔊Text to speech🗣️Voice cloning🗒Transcription
  NewAudio

  Released 1mo ago
• Gen 3 Nemotron
  Nemotron Labs Audex 30B A3B

  By NVIDIA

  Nemotron-Labs-Audex-30B-A3B is a unified audio-text large language model from NVIDIA built on a 30B parameter Mixture-of-Experts backbone with 3B active parameters. It extends a strong text-only reasoning model with an audio encoder and discrete audio token vocabulary, enabling audio understanding, speech recognition, speech translation, text-to-speech, audio generation, and speech-to-speech interaction while preserving the backbone's reasoning, knowledge, and long-context abilities.

  🗣️Speech to speech🔊Text to speech🎧Audio translation🗒Transcription🔍Advanced reasoning
  NewMultimodal

  Released 1mo ago
• Gen 4
  Higgs Audio v3 TTS

  By Boson AI

  Higgs Audio v3 TTS is Boson AI’s text-to-speech model for expressive conversational voice agents across 100+ languages with zero-shot voice cloning and inline speech controls.

  🔊Text to speech🗣️Voice cloning🎙️Voiceovers🗣Dialogue generation
  NewAudio

  Released 1mo ago
• Gen 4
  Miso TTS 8B

  By Miso Labs

  MisoTTS is Miso Labs’ open-weight 8B text-and-audio-conditioned speech generation model for expressive, context-aware, emotive TTS and dialogue voice output.

  🔊Text to speech🗣️Voice cloning🎙️Voiceovers🗣Dialogue generation
  NewAudio

  Released 1mo ago
• Gen 3 MAI
  MAI Voice 2 Flash

  By Microsoft

  MAI-Voice-2-Flash is Microsoft AI’s upcoming lower-cost, ultra-efficient variant of MAI-Voice-2 for speech generation.

  🔊Text to speech🗣️Voice cloning🎙️Voiceovers
  NewMultimodal

  Released 1mo ago
• Gen 3 MAI
  MAI Voice 2

  By Microsoft

  MAI-Voice-2 is Microsoft AI’s speech generation model for natural-sounding voice output across 15 languages with short-sample voice adaptation.

  🔊Text to speech🗣️Voice cloning🎙️Voiceovers🎧Audiobooks
  NewMultimodal

  Released 1mo ago
• Gen 4
  Phonon

  By Gradium

  Phonon is Gradium’s private-beta 100M-parameter on-device text-to-speech model for low-latency, offline, privacy-sensitive voice generation.

  🔊Text to speech🗣️Voice cloning
  NewVideo

  Released 2mo ago
• Gen 3
  Realtime TTS 2

  By Inworld AI

  Realtime TTS-2 is Inworld AI’s realtime conversational text-to-speech model. It is built for live voice interaction rather than narration, with conversational awareness from prior audio turns, natural-language voice direction, crosslingual voice identity across 100+ languages, and prompt-based voice design.

  🔊Text to speech🗣️Voice cloning🎙️Voiceovers
  NewMultimodal

  Released 2mo ago
• Gen 3
  Sonic 3.5

  By Cartesia

  Sonic 3.5 is Cartesia’s fastest and most natural text-to-speech model, built for low-latency conversational voice generation across 42 languages.

  🔊Text to speech🗣️Voice cloning🎙️Voiceovers🗣Dialogue generation
  NewMultimodal

  Released 2mo ago
• Gen 3
  p video avatar

  By Pruna AI

  p-video-avatar is Pruna’s talking-head video generation model for creating speaking avatar videos from a single portrait image. It takes either a text script or an audio file, then generates a realistic head-and-shoulders speaking video, with support for multiple voices, languages, and 720p or 1080p output.

  🎥Video avatars🔊Text to speech🎤Lip sync videos🎨Portrait animation
  NewMultimodal

  Released 2mo ago
• Gen 3
  sarashina 2.2 tts

  By SB Intuitions

  sarashina2.2-tts is SB Intuitions’ Japanese-centric large-language-model-based text-to-speech system. It supports Japanese and English, is designed for high pronunciation accuracy, naturalness, and stability across diverse speaking styles, and includes zero-shot voice generation.

  🔊Text to speech🗣️Voice cloning🎙️Voiceovers
  Multimodal

  Released 3mo ago
• Gen 3
  Soniox Text to Speech

  By Soniox

  Soniox Text-to-Speech is Soniox’s multilingual TTS model and API for precise, low-latency speech generation. It is built for production voice systems, supports 60+ languages, and emphasizes accurate pronunciation, faithful reading of structured text like emails and phone numbers, natural code-switching, and streaming output for real-time voice apps.

  🔊Text to speech🎙️Voiceovers🎤Voice agents
  Multimodal

  Released 3mo ago
• Gen 3
  OpenRouter Text to Speech

  By OpenRouter

  OpenRouter TTS is OpenRouter’s unified text-to-speech interface for accessing multiple speech-generation models through one API layer. It standardizes voice generation across providers, supporting streaming audio output, customizable voices, and multimodal workflows without provider-specific integration complexity.

  🔊Text to speech🎙Voice chatting🎙️Voiceovers
  Multimodal

  Released 3mo ago
• Gen 3
  StepAudio 2.5 TTS

  By StepFun

  StepAudio 2.5 TTS is StepFun’s contextual text-to-speech model with performance-oriented vocal control. It combines global and inline context guidance with zero-shot voice cloning so generated speech can follow broader style instructions as well as local delivery details, rather than just reading text flatly.

  🔊Text to speech🗣️Voice cloning🗣️Dialect simulation
  Multimodal

  Released 3mo ago
• Gen 4 MAI
  MAI Voice 1

  By Microsoft

  MAI-Voice-1 is Microsoft’s top-tier text-to-speech model for natural, expressive voice generation. It is built to preserve clarity, intent, speaker identity, emotional nuance, and pacing across long-form speech, and supports custom voice creation from only a few seconds of audio. Microsoft positions it for voice experiences, voice agents, and expressive spoken content at high speed and low cost.

  🔊Text to speech🎤Voice changing🗣️Voice cloning🎧Audiobooks🔊Advanced audio generation
  Audio

  Released 3mo ago
• Gen 3 MOSS
  MOSS TTS Nano

  By OpenMOSS

  MOSS-TTS-Nano is an open-source multilingual tiny speech generation model from MOSI.AI and OpenMOSS. With only 0.1B parameters, it is built for real-time TTS, can run directly on CPU without a GPU, and keeps deployment simple enough for local demos, web serving, and lightweight product integration.

  🔊Text to speech🗣️Voice cloning🌐Multilingual communication
  Multimodal

  Released 3mo ago
• Gen 4
  Pocket TTS Spokenword

  By danneauxs

  Pocket-TTS-Spokenword is an enhanced version of Kyutai’s Pocket TTS built for emotionally expressive audiobook generation from plain text. It adds AI emotion analysis, smart text chunking, voice adaptation, and voice cloning, while staying lightweight enough to run on CPU-only systems without requiring a GPU.

  🔊Text to speech🗣️Voice cloning🎧Audiobooks
  Audio

  Released 3mo ago
• Gen 3
  VoxCPM2

  By OpenBMB

  VoxCPM2 is OpenBMB’s open-source tokenizer-free multilingual text-to-speech model for natural speech generation, voice design, and controllable voice cloning. It is a 2B-parameter model trained on over 2 million hours of speech, supports 30 languages, and produces 48 kHz studio-quality audio with real-time streaming capability.

  🔊Text to speech🎤Voice changing🗣️Voice cloning🔊Audio
  Multimodal

  Released 3mo ago
• Gen 3
  OmniVoice

  By Xiaomi

  OmniVoice is a multilingual zero-shot text-to-speech model built for voice cloning, voice design, and general speech synthesis at massive language scale. It supports more than 600 languages, uses a diffusion language model-style architecture, and is positioned for high-quality speech generation with fast inference.

  🔊Text to speech🎤Voice changing🗣️Voice cloning🌐Multilingual communication
  Multimodal

  Released 3mo ago
• Gen 4 LongCat
  LongCat AudioDiT 3.5B

  By Meituan

  LongCat-AudioDiT-3.5B is Meituan LongCat’s diffusion-based text-to-speech model built directly in waveform latent space rather than mel-spectrogram space. It is designed for high-fidelity speech generation and zero-shot voice cloning, supports Chinese and English, and is positioned as a top-performing open model on the Seed benchmark for speaker similarity and intelligibility.

  🔊Text to speech🗣️Voice cloning🔊Voice enhancement
  Audio

  Released 3mo ago
• Gen 3
  Voxtral TTS

  By Mistral AI

  Voxtral TTS is Mistral’s new open-source text-to-speech model for building voice agents and enterprise speech applications. According to TechCrunch, it supports 9 languages, can clone a voice from under 5 seconds of audio, preserves accents and speaking style, and is optimized for real-time use on edge devices like phones, laptops, and wearables.

  🔊Text to speech🗣️Voice cloning🎙️Voiceovers🎤Voice agents
  Audio

  Released 4mo ago
• Gen 4
  Lightning v3

  By Smallest AI

  Lightning is Smallest.ai’s low-latency text-to-speech system for real-time voice agents, voiceovers, and voice cloning.

  🔊Text to speech🗣️Voice cloning🎙️Voiceovers🎤Voice agents
  Audio

  Released 4mo ago
• Gen 4 MOSS
  MOSS TTS Local Transformer v1.5

  By OpenMOSS

  MOSS-TTS-Local-Transformer-v1.5 is a 5B-parameter text-to-speech model supporting 31 languages with zero-shot voice cloning, long-form speech generation, token-level duration control, Pinyin/IPA pronunciation control, code-switching, and 48 kHz stereo audio output via MOSS-Audio-Tokenizer-v2.

  🔊Text to speech🗣️Voice cloning🎙️Voiceovers🎧Audiobooks
  Audio

  Released 4mo ago
• Gen 3 MiMo
  Xiaomi MiMo V2 TTS

  By Xiaomi

  MiMo-V2-TTS is Xiaomi’s large-scale speech synthesis model built for expressive agent voice, aiming for natural, emotionally aware speech.

  🔊Text to speech🎤Voice changing🎙️Voiceovers🎤Singing
  Audio

  Released 4mo ago
• Gen 4
  TADA 1B

  By Hume AI

  TADA-1B is a unified speech-language model checkpoint that aligns text tokens and speech representations 1-to-1 for fast, reliable text-to-speech generation.

  🔊Text to speech🗣️Voice cloning🗣️Speech to speech🎙️Voiceovers
  Audio

  Released 4mo ago
• Gen 3
  TADA 3B ML

  By Hume AI

  TADA-3B-ml is a multilingual TADA checkpoint built for fast, reliable speech generation using the same 1-to-1 text-acoustic alignment framework.

  🔊Text to speech🗣️Voice cloning🗣️Speech to speech🎙️Voiceovers🌐Multilingual communication
  Multimodal

  Released 4mo ago
• Gen 3
  Qwen3 TTS

  By Alibaba

  Qwen3-TTS is a speech generation model family designed for high-quality, human-like TTS with voice cloning and natural-language control over voice style.

  🔊Text to speech🎤Voice changing🗣️Voice cloning🎙️Voiceovers
  Audio

  Released 5mo ago
• Gen 3 GPT
  GPT Realtime 1.5

  By OpenAI

  gpt-realtime-1.5 is OpenAI’s flagship real-time voice model for audio-in, audio-out use cases like voice agents and customer support, with support for text, audio, and image inputs and text and audio outputs.

  🎙Voice chatting🔊Text to speech🎤Voice assistants
  Multimodal

  Released 5mo ago
• Gen 2
  CSM

  By Sesame AI Labs

  Conversational speech generation model that generates audio codes from text and audio inputs for dialogue style speech output.

  🔊Text to speech
  Coding

  Released 5mo ago
• Gen 3 Seed
  Seed 2.0

  By ByteDance

  Seed 2.0 is described publicly only as a new ByteDance Seed language model for Doubao, but there is not yet any reliable, detailed public technical description of its architecture, context length, or training data that I can quote.

  🎬Video editing🔊Text to speech📰News analysis
  Multimodal

  Released 5mo ago
• Gen 3 Seedream
  Seedream 5.0 Lite

  By ByteDance

  Seedream 5.0 Lite is ByteDance Seed's multimodal image generator with deep reasoning and built in web search, built for precise text to image and image editing that follow complex, real time instructions with tight layout and style control.

  🖼️Image generation🔊Text to speech🖌️Image editing🖼️Logos
  Image

  Released 5mo ago
• Gen 4
  Zyphra

  By Zyphra AI

  Zonos-v0.1 is Zyphra’s open-weight text-to-speech family, two 1.6B models trained on 200k+ hours of multilingual speech, offering expressive, real-time TTS and high-quality voice cloning.

  🔊Text to speech🗣️Voice cloning
  Audio

  Released 5mo ago
• Gen 4
  LuxTTS

  By ysharma3501

  ZipVoice-based voice cloning TTS that generates 48 kHz speech at up to 150x real time, fitting in about 1 GB VRAM for local, high quality synthesis

  🔊Text to speech🗣️Voice cloning
  Audio

  Released 5mo ago
• Gen 3
  MOVA

  By OpenMOSS

  Open source foundation model that jointly generates video and audio in one pass, achieving tightly synchronized lip movements and environment-aware sound effects.

  🎥Videos🔊Text to speech🎵Music🎬Animations
  Video

  Released 5mo ago
• Gen 4 Qwen
  Qwen3 ForcedAligner 0.6B

  By Alibaba

  Multilingual forced alignment model that aligns speech and transcripts in 11 languages, predicting timestamps for arbitrary units in up to 5 minutes of audio with accuracy surpassing previous end-to-end aligners.

  🗒Transcription🔊Text to speech🌐Text translation🔍SEO content
  Audio

  Released 5mo ago
• Gen 4 MiniMax
  Minimax Speech 2.8 Turbo

  By MiniMax

  Latency-optimized sibling of Speech-2.8-HD, trading a bit of ultimate fidelity for faster, cheaper generation while keeping multilingual, emotional, voice-cloning strengths.

  🔊Text to speech🗣️Voice cloning
  Audio

  Released 6mo ago
• Gen 4 MiniMax
  Minimax Speech 2.8 HD

  By MiniMax

  High-definition MiniMax TTS model focused on studio-grade, multilingual speech, rich emotion control, interjections and voice cloning for premium voiceovers and production audio.

  🔊Text to speech🗣️Voice cloning
  Audio

  Released 6mo ago
• Gen 2
  D4RT

  By Google

  D4RT is DeepMind’s unified 4D scene reconstruction and tracking model that turns ordinary videos into a fast, queryable representation of 3D geometry and motion, solving tracking, depth and pose up to hundreds of times faster than prior work.

  🌍3D images🔊Text to speech🎮Game creation🎬Video editing
  Coding

  Released 6mo ago
• Gen 4 Qwen
  Qwen3 TTS 12Hz 1.7B CustomVoice

  By Alibaba

  Qwen’s open text-to-speech model supporting multilingual speech generation with custom voice capability.

  🔊Text to speech🗣️Voice cloning
  Audio

  Released 6mo ago
• Gen 3
  Manzano

  By Apple

  Manzano is Apple’s unified multimodal model that shares a hybrid vision tokenizer for both image understanding and text-to-image generation, using one autoregressive LLM plus a diffusion decoder to reach state-of-the-art unified performance.

  🖼️Image generation🔊Text to speech🔍SEO content🎮Game creation
  Multimodal

  Released 6mo ago
• Gen 7
  Chroma 1.0

  By Flash Labs

  FlashLabs Chroma 1.0 is a real-time spoken dialogue model that interleaves text and audio tokens to enable sub-second, end-to-end conversations with personalized voice cloning and high speaker similarity.

  🔊Text to speech🗣️Voice cloning
  Text

  Released 6mo ago
• Gen 4 FLUX
  FLUX.2 [klein] 4B

  By Black Forest Labs

  FLUX.2-klein-4B is Black Forest Labs’ 4B-parameter rectified-flow image model, unifying fast text-to-image and image-editing with multi-reference support, distilled for sub-second generation on consumer GPUs under Apache 2.0.

  🖼️Image generation🔊Text to speech🔍SEO content🖌️Image editing
  Image

  Released 6mo ago
• Gen 7 Seed
  Seed Prover 1.5

  By ByteDance

  Seed-Prover 1.5 is ByteDance Seed’s formal theorem-proving model for Lean, trained with agentic RL and test-time scaling to solve most undergraduate and many graduate-level competition problems.

  📚Academic research🔊Text to speech🎨NFT art
  Text

  Released 6mo ago