Voice & audio

Voices, transcription, voice agents, music and sound. 5 tasks, 260 listed options.

Tier: how close to the best result on the task's main public benchmark
Outlined: a tool inherits the tier of the best model it offers
No public result yet
Cloud · Local: runs on your own hardware

Text to speech

Natural voices for narration, voiceovers and voice cloning.

Best overallSEleven v4 Turbo
Best valueASimba 3.2
Best free startSElevenLabs
Best local (≤24 GB GPU)BBreeze TTS 2

Models

Cloud

Eleven v4 TurboElevenLabs · $40/M characters$S

Eleven v4ElevenLabs · $80/M characters S

Qwen-Audio-3.1-TTS-PlusAlibaba (Qwen) · $19/M characters$A

Sonic 3.6Cartesia · $49/M characters A

Gemini 3.8 Flash TTSGoogle · $16/M characters$A

Realtime TTS-2Inworld AI · $21/M characters A

Gemini 3.8 Flash-Lite TTSGoogle · $11/M characters$A

Simba 3.2Speechify · $6.58/M characters$B

Luna TTSVUI Labs · $15/M characters B

StepAudio 2.5 TTS (Aug 2026)StepFun · $85/M characters B

Fish Audio S2.1 ProFish Audio · $15/M characters B

Soniox TTS Real-Time v2Soniox · $14/M characters B

Gradium TTS (Sep 2026)Gradium · $48/M characters B

Speech 2.8 HDMiniMax · $100/M characters B

Lightning V3.1 Pro (Jul 2026)Smallest.ai · $20/M characters B

Falcon 2Murf AI · $10/M characters C

SpaceXAI TTS (AA name)xAI · $15/M characters C

Async Flash v1.5Async · $10/M characters C

Azure HD 2.5Microsoft · $22/M characters C

Async Pro v1.0Async · $20/M characters C

Maya 2Maya Research C

Polly GenerativeAmazon · $30/M characters C

CodaRime · $50/M characters C

Octave 2 (preview)Hume AI · $88/M characters C

MiMo-V2.5-TTSXiaomi C

Flux TTSDeepgram · $45/M characters C

Bland Speech v3Bland AI · $40/M characters C

MAI-Voice-1Microsoft AI · $22/M characters C

Kugel 3KugelAudio –

GPT-4o Mini TTS (2025-12-15)OpenAI –

Mist v3Rime –

SSFM 3.0 (ssfm-v30)Neosapience (Typecast)

Local

Fish Audio Drama 3 (Preview)Fish Audio · ~48 GB VRAM$B

Breeze TTS 2BreezeBlue · ~12 GB VRAM$C

Step-Audio-EditXStepFun · ~8 GB VRAM$C

Voxtral TTS (4B, 2603)Mistral AI · ~10 GB VRAM C

Kokoro-82M v1.0hexgrad · ~1 GB VRAM$C

Magpie-Multilingual 357M (Feb 2026)NVIDIA · ~1 GB VRAM C

Maya1Maya Research · ~8 GB VRAM C

Higgs TTS 3 (4B)Boson AI · ~12 GB VRAM C

VibeVoice-1.5BMicrosoft · ~8 GB VRAM –

Chatterbox Multilingual V3Resemble AI · ~2 GB VRAM –

Dia2-2BNari Labs · ~6 GB VRAM –

F5-TTS v1SWivid (Shanghai Jiao Tong University) · Open weights –

Fun-CosyVoice3-0.5B-2512Alibaba (FunAudioLLM) · ~2 GB VRAM –

IndexTTS-2.5bilibili (IndexTeam) · Open weights –

Kitten TTS 2KittenML · Open weights –

MOSS-TTS v1.5OpenMOSS (Fudan University / MOSI) · ~19 GB VRAM –

NeuTTS AirNeuphonic · ~2 GB VRAM –

OmniVoicek2-fsa (Next-gen Kaldi) · ~2 GB VRAM –

Orpheus 3B 0.1Canopy Labs · ~9 GB VRAM –

Pocket TTSKyutai · Open weights –

Qwen3-TTS-12Hz-1.7BAlibaba (Qwen) · ~6 GB VRAM –

VoxCPM2OpenBMB · ~6 GB VRAM –

ZONOS2Zyphra · Open weights

Apps

Cloud

ElevenLabsElevenLabs · Free tier · $6/mo$A

SpeechifySpeechify · Free tier · $29/mo A

Fish AudioFish Audio · Free tier · $15/mo B

MurfMurf AI · Free tier · $29/mo –

DescriptDescript · Free tier · $24/mo –

NarakeetNarakeet –

NaturalReaderNaturalSoft · Free tier –

TypecastNeosapience · Free tier · $5/mo –

WellSaidWellSaid Labs · Free tier · $19/mo

Local

VoiceboxJamie Pine (open source) · Free$–

ebook2audiobookDrewThomasson (open source) · Free

Local servers & APIs

Cloud

Nothing tracked yet

Local

C sherpa-onnxk2-fsa (Next-gen Kaldi) · Free$C

Kokoro-FastAPIremsky (open source) · Free$C

Speachesspeaches-ai (open source) · Free$C

TTS WebUIrsxdalv (open source) · Free$C

NVIDIA NeMo SpeechNVIDIA · Free –

PiperOpen Home Foundation · Free –

AllTalk TTSerew123 (open source) · Free

Full text to speech map, chart and table →

Transcription

Turn speech into text: transcripts, captions, dictation.

Best overallSFun-Realtime-ASR (preview)
Best valueAModulate STT (batch)
Best free startSElevenLabs
Best local (≤24 GB GPU)AQwen3-ASR-1.7B

Models

Cloud

Fun-Realtime-ASR (preview)Alibaba S

MAI-Transcribe-2 (preview)Microsoft AI · $0.0017/min of audio$S

Scribe v2ElevenLabs · $0.0037/min of audio S

Grok Voice Transcribe 2.0xAI · $0.0017/min of audio S

StepAudio 3 ASRStepFun · $0.0067/min of audio S

Gemini 3.5 TranscribeGoogle · $0.0050/min of audio A

Ink-2Cartesia · $0.0040/min of audio A

Scribe V2 ProZoom · $0.0033/min of audio A

Saaras v4Sarvam AI · $0.0052/min of audio A

Universal-3.6 Pro StreamingAssemblyAI · $0.0075/min of audio A

Gemini 3.1 Pro (Preview)Google · $0.02/min of audio A

Universal-3.5 ProAssemblyAI · $0.0035/min of audio A

Solaria-3Gladia · $0.01/min of audio A

GPT-TranscribeOpenAI · $0.0045/min of audio A

PulseSmallest.ai · $0.0050/min of audio A

Resonant-1Reson8 · $0.0036/min of audio A

Muse Voice TranscribeMeta · $0.0030/min of audio A

Qwen3.5-Omni-PlusAlibaba (Qwen)A

Voxtral Mini Transcribe 2 (v26.02)Mistral AI · $0.0030/min of audio A

Soniox v5 (stt-rt-v5 / async)Soniox · $0.0017/min of audio A

Inworld STT 1Inworld AI · $0.0025/min of audio A

Amazon TranscribeAmazon · $0.0060/min of audio A

Modulate STT (batch)Modulate · $0.00042/min of audio$B

Chirp 3Google · $0.02/min of audio B

Nova-3Deepgram · $0.0043/min of audio B

Melia 1Speechmatics · $0.0040/min of audio C

Flux (English / Multilingual)Deepgram · $0.0065/min of audio C

Gradium Speech-to-TextGradium · $0.01/min of audio

Local

Qwen3-ASR-1.7BAlibaba (Qwen) · ~6 GB VRAM$A

Voxtral Small (24B-2507)Mistral AI · ~55 GB VRAM A

Higgs Audio v3 STTBoson AI · ~6 GB VRAM A

Granite Speech 4.1 2BIBM · ~6 GB VRAM A

MOSS-Transcribe-DiarizeOpenMOSS (Fudan University / MOSI) · ~2 GB VRAM$A

distil-large-v3.5Hugging Face · ~2 GB VRAM A

Whisper large-v3OpenAI · ~10 GB VRAM B

Canary-Qwen-2.5BNVIDIA · ~6 GB VRAM B

Fun-ASR-Nano-2512Alibaba (FunAudioLLM) · ~2 GB VRAM B

Parakeet TDT 0.6B v3NVIDIA · ~2 GB VRAM B

cohere-transcribe-03-2026Cohere · ~6 GB VRAM B

omniASR-LLM-7B-v2Meta (FAIR) · ~16 GB VRAM B

VibeVoice-ASRMicrosoft · ~19 GB VRAM B

Voxtral Mini 4B Realtime (2602)Mistral AI · ~10 GB VRAM C

Nemotron 3.5 ASR Streaming 0.6BNVIDIA · ~2 GB VRAM –

MiMo-V2.5-ASRXiaomi · ~17 GB VRAM –

Moonshine StreamingMoonshine AI (formerly Useful Sensors) · Open weights –

stt-2.6b-enKyutai · ~6 GB VRAM

Apps

Cloud

ElevenLabsElevenLabs · Free tier · $6/mo$A

TurboScribeTurboScribe · Free tier · $20/mo –

Aqua VoiceAqua Voice · Free tier · $8/mo –

DescriptDescript · Free tier · $24/mo –

NottaNotta · Free tier · $8.17/mo –

Otter.aiOtter.ai · Free tier · $17/mo –

RevRev.com · Free tier · $30/mo –

SonixSonix · Free tier · $25/mo –

WillowWillow · Free tier · $12/mo –

Wispr FlowWispr · Free tier · $15/mo

Local

VoiceboxJamie Pine (open source) · Free$A

HandyCJ Pais (open source) · Free$A

BuzzChidi Williams (open source) · Free$A Vibethewh1teagle (open source) · Free$A

VoiceInkPrakash Joshi Pax (open source)A Whishperpluja (open source) · Free$A

AikoSindre Sorhus · Free tier A

MacWhisperJordi Bruin (Good Snooze) · Free tier A

SpokenlySpokenly · Free$A

SuperwhisperSuperwhisper · Free tier · $8.49/mo

Engines & APIs

Cloud

Nothing tracked yet

Local

whisper.cppggml-org (open source) · Free$A

faster-whisperSYSTRAN (open source) · Free$A

WhisperXMax Bain (open source) · Free$A sherpa-onnxk2-fsa (Next-gen Kaldi) · Free$A

WhisperKitArgmax · Free$A

WhisperLiveCollabora · Free$A

Speachesspeaches-ai (open source) · Free$B

NVIDIA NeMo SpeechNVIDIA · Free –

VoskAlpha Cephei · Free –

Moonshine VoiceMoonshine AI · Free

Full transcription map, chart and table →

Voice agents

Talk with AI in real time, or build agents that answer calls.

Best overallSGemini 3.8 Live
Best valueSGemini 3.8 Live
Best free startSCartesia Line
Best local (≤24 GB GPU)No qualifying option yet

Models

Cloud

Gemini 3.8 LiveGoogle · $0.01/min of audio$S

Grok Voice 3.0xAI · $0.08/min of audio A

GPT-Live 1OpenAI · $0.05/min of audio B

GPT-Realtime-2.1OpenAI · $0.19/min of audio C

Qwen3.5-Omni-PlusAlibaba (Qwen)C

GPT-Realtime-2.1 MiniOpenAI · $0.08/min of audio C

Qwen Audio 3.0 Realtime PlusAlibaba (Qwen) · $0.07/min of audio –

Higgs RealtimeBoson AI –

Nova 2.5 SonicAmazon –

OpalDeepslate · $0.11/min of audio –

StepAudio 3 RealtimeStepFun

Local

LFM2.5-Audio-1.5BLiquid AI · ~4 GB VRAM –

MiniCPM-o 4.5OpenBMB · ~21 GB VRAM –

Moshi (moshiko)Kyutai · ~24 GB VRAM –

NVIDIA-NemotronLabs-VoiceChat-11BNVIDIA · ~25 GB VRAM –

PersonaPlex 7B v1NVIDIA · ~24 GB VRAM –

Qwen3-Omni-30B-A3B-InstructAlibaba (Qwen) · ~80 GB VRAM –

Raon-SpeechChat-9BKRAFTON · ~22 GB VRAM –

Step-Audio R1.1StepFun · ~74 GB VRAM

Apps

Cloud

GeminiGoogle · Free tier · $4.99/mo$S

GrokSpaceXSI (formerly xAI) · Free tier · $8/mo A

ChatGPTOpenAI · Free tier · $20/mo –

ClaudeAnthropic · Free tier · $20/mo –

Meta AIMeta · Free tier – Microsoft CopilotMicrosoft · Free tier · $9.99/mo –

SesameSesame AI

Local

Nothing tracked yet

Voice-agent platforms

Cloud

Cartesia LineCartesia · Free tier · $5/mo$S

ElevenLabsElevenLabs · Free tier · $6/mo B

Deepgram Voice Agent APIDeepgram · Free tier B

Retell AIRetell AI · Free tier –

BlandBland AI · Free tier –

Millis AIMillis AI –

PolyAIPolyAI –

SynthflowSynthflow AI –

Telnyx Voice AI AgentsTelnyx –

ThoughtlyThoughtly –

UltravoxUltravox (Fixie) · Free tier –

VapiVapi · Free tier –

VoiceflowVoiceflow

Local

BolnaBolna AI · Free tier

Frameworks

Cloud

Nothing tracked yet

Local

PipecatDaily (open source) · Free –

LiveKit AgentsLiveKit · Free –

speech-to-speech (Hugging Face)Hugging Face · Free –

TEN FrameworkTEN (Agora) · Free –

UnmuteKyutai · Free

Full voice agents map, chart and table →

Music generation

Songs, instrumentals and stems from a prompt.

Best overallSSuno v6
Best valueNo qualifying option yet
Best free startSSuno
Best local (≤24 GB GPU)BMiniMax Music 3.0

Models

Cloud

Suno v6-miniSuno S

Mureka V9.5Kunlun Tech (Mureka)A

Lyria 3.5Google DeepMind B

Eleven Music v2.5ElevenLabs · $0.15/min of audio$B

Stable Audio 3 LargeStability AI B

Melodia v3Treblo (formerly Sonauto)–

HappyShrimp 1.0 (beta)Alibaba (Alibaba Token Hub)

Local

MiniMax Music 3.0MiniMax · ~24 GB VRAM$B

Stable Audio 3 MediumStability AI · Open weights C

MusicGenMeta · Open weights –

ACE-Step 1.5ACE Studio and StepFun · ~8 GB VRAM –

AudioX-TurboHKUST (HKUSTAudio) and NoizAI · Open weights –

DiffRhythm 2ASLP-lab (Northwestern Polytechnical University) · Open weights –

HeartMuLa-oss-3B-happy-new-yearHeartMuLa Team · Open weights –

Magenta RealTime 2Google DeepMind · Open weights –

YuE2 (3B)M-A-P (Multimodal Art Projection) · ~24 GB VRAM

Apps

Local

WanGP (Wan2GP)deepbeepmeep (community project) · Free$–

ACE-Step UIfspecii (community) · Free –

yovoiceleemysw (community) · Free

Workflow platforms

Cloud

Nothing tracked yet

Local

ComfyUIComfy Org · Free$

Runtimes & APIs

Cloud

Nothing tracked yet

Local

audio.cpp0xShug0 (community) · Free$C

TTS WebUIrsxdalv (open source) · Free

Full music generation map, chart and table →

Sound effects

Sound effects and foley from text or video.

No public leaderboard covers this task yet, so its options are mapped but not ranked.

Models

Cloud

Mirelo SFX 2.0Mirelo AI –

Sound Effects v2 (eleven_text_to_sound_v2)ElevenLabs · $0.12/min of audio

Local

AudioX-TurboHKUST (HKUSTAudio) and NoizAI · Open weights –

ControlFoleyXiaomi (xiaomi-research) · Open weights –

Dasheng-AudioGenXiaomi (MiLM Plus / mispeech) · Open weights –

HunyuanVideo-Foley (XXL and XL)Tencent Hunyuan · ~8 GB VRAM –

MMAudio (large_44k_v2)UIUC and Sony AI · ~6 GB VRAM –

MOSS-SoundEffect v2.0OpenMOSS team and MOSI.AI · Open weights – PrismAudioAlibaba (FunAudioLLM) · Open weights –

Stable Audio 3 Small SFXStability AI · Open weights –

Woosh v1.0.1 (Flow, DFlow, VFlow, DVFlow)Sony AI · Open weights

Apps

Cloud

Adobe FireflyAdobe · $9.99/mo –

Beatoven.aiBeatoven.ai · Free tier · $10/mo –

ElevenLabsElevenLabs · Free tier · $6/mo –

MireloMirelo AI · Free tier –

OptimizerAIOptimizerAI · Free tier –

SFX EngineSFX Engine · Free tier · $10/mo –

WubbleWubble AI · Free tier · $4.99/mo

Local

WanGP (Wan2GP)deepbeepmeep (community project) · Free

Workflow platforms

Cloud

Nothing tracked yet

Local

ComfyUIComfy Org · Free

Runtimes & APIs

Cloud

Nothing tracked yet

Local

audio.cpp0xShug0 (community) · Free

Full sound effects map, chart and table →