whatmodel.sh · find the best text-to-speech models you can run

Best text-to-speech models for a 8GB graphics card

A leaderboard of the best text-to-speech models for a 8GB graphics card (with 32GB of system RAM). The best one that runs on it right now is dots.tts MF 2B (MeanFlow, NFE 4), followed by FireRedTTS3 (base) and LongCat-AudioDiT-3.5B. Models are ranked by one score that combines 101 published benchmarks. Models that run on it show the download that fits and the command to run it.

149 models · scores combine 101 benchmarks · updated 2026-10-11 02:20 UTChow it works · find a leaderboard · benchmark your machine

task[all tasks][voice cloning]

Models for 32GB RAM · 8GB graphics card → ~8GB fast · reset

+ more

set[all sets][Vapi Humanness · Bradley–Terry rating, 2026-09-25 snapshot · 148][VITS 2021 LJ Speech human MOS · 5][Grad-TTS 2021 LJ Speech human MOS · 4]

Filter the model list while keeping overall scores. Models without a shared score scale are not yet calibrated; select a set to see its own scores.

Detect your exact chip or graphics card:

curl -G whatmodel.sh/tts -d max=0 --data-urlencode "mac=$(sysctl -n hw.memsize machdep.cpu.brand_string)"Mac

curl -G whatmodel.sh/tts -d max=0 --data-urlencode "ram=$(awk '/MemT/{print $2}' /proc/meminfo)" --data-urlencode "nv=$(nvidia-smi --query-gpu=name,memory.total --format=csv,noheader 2>/dev/null)"Linux

curl.exe -G whatmodel.sh/tts -d max=0 --data-urlencode "ram=$((gcim Win32_ComputerSystem).TotalPhysicalMemory)" --data-urlencode "nv=$(nvidia-smi --query-gpu=name,memory.total --format=csv,noheader 2>$null)"Windows

$ terminal

ram[8][16][24][32][48][64][96][128][192][256]GB

gpu[apple silicon][graphics card][none]

vram[1][2][4][8][12][16][24][32][48][64][80][96][128][160]GB

price[local][$1][$2][$5][$10][$20][$50][any]max per M chars

detecting your OS…

Score vs API price

     │                                                   OpenAudio S1 ◆● Voxtral TTS (4B)
  60 ┤ · · · · · ··· · · · · · · · · · · ·Simba 3.2·◆─────────◇───────◇· · ◇ · ·◇· ◇ · · · · ◇ ·◇· · · · ·◇· ·
     │   ●──────────────────────────────────────────┘                    ◇      ∘  ◇    ◇    ◇    ◇       ◇
  40 ┤ · · · · · ··· · · · · · · · · · · · · · · · · · · · · ··· · · ·∘◇ · · · · · · · · · · · · · · · · ··· ·
     │                                                 ∘                                 ◇            ◇
     └────────────┬───────────────────────────────────────────┬───────────────────────────────────────────┬───
                 $1                                          $10                                        $100

#	model	score↓	where↓	speed↓	RAM↓	get it	sources

#	model	score↓	where↓	RAM↓	get it	sources
4	Voxtral TTS (4B) Mistral	71±4†	api $16.00/M chars +local	~10GB	docs.mistral.ai ↗
5	OpenAudio S1 Fish Audio	71±8*†	api $15.00/M chars		docs.fish.audio ↗
7	Qwen-Audio-3.0-TTS-Plus Alibaba	69±2†	api $20.00/M chars		www.alibabacloud.com ↗
8	Gemini 3.1 Flash TTS Preview Google	69±2	api $50.00/M chars		fal.ai ↗	+1
9	dots.tts MF 2B (MeanFlow, NFE 4) · bf16 (dots.tts package CLI) dots.tts	68±3†	local	~5.2GB	how to run ›
11	Grok TTS xAI	68±3	api $15.00/M chars		docs.x.ai ↗
12	FireRedTTS3 (base) · bf16 (FireRedTTS3 repo) FireRedTeam (Xiaohongshu)	67±3	local (slow)	~12GB	how to run ›
13	Simba 3.2 Speechify	67±3*	api $6.00/M chars		speechify.ai ↗
15	MiniMax Speech 2.8 Turbo MiniMax	66±3*	api $60.00/M chars		platform.minimax.io ↗
16	LongCat-AudioDiT-3.5B · fp32 weights (official LongCat-AudioDiT repo, transformers>=5.3) Meituan	66±2	local (slow)	~15GB	how to run ›
17	LongCat-AudioDiT-1B · fp32 weights (official LongCat-AudioDiT repo, transformers>=5.3) Meituan	66±4†	local	~5.7GB	how to run ›
18	OmniVoice · fp32 k2-fsa (Xiaomi)	66±2	local	~3.3GB	how to run ›
19	MiniMax Speech 2.8 HD MiniMax	66±2	api $100.00/M chars		platform.minimax.io ↗
20	dots.tts SOAR 2B · bf16 (dots.tts package CLI) dots.tts	65±2	local	~5.2GB	how to run ›
21	Eleven v3 ElevenLabs	65±1	api $100.00/M chars		elevenlabs.io ↗	+2
22	Chirp 3: HD Google	65±4*	api $30.00/M chars		cloud.google.com ↗
23	Higgs Audio V3 TTS Boson AI	65±3	api $15.00/M chars +local	~10GB	docs.boson.ai ↗
25	Qwen3-TTS 12Hz 1.7B Base · bf16 (qwen-tts) Alibaba (Qwen)	65±1	local	~4.5GB	how to run ›	+3
26	VoxCPM 0.5B · bf16 OpenBMB	64±1	local	~5GB	how to run ›
27	VoxCPM2 · bf16 (voxcpm) OpenBMB	64±1	local	~5GB	how to run ›
28	Fun-CosyVoice3-0.5B-2512 · fp32 (CosyVoice repo) Alibaba (FunAudioLLM)	64±1	local	~5.4GB	how to run ›
29	Wayu-Paxa-TTS-Edge · fp32 (wayu-tts package) Wayu Research / Paxa Labs	64±4†	local	~0.3GB	how to run ›
30	Eleven Turbo v2.5 ElevenLabs	64±4*	api $50.00/M chars		elevenlabs.io ↗
31	Fish Audio S2.1 Pro Fish Audio	63±3*	api $15.00/M chars		docs.fish.audio ↗
32	IndexTTS-2 · fp32 (index-tts repo) IndexTeam (bilibili)	63±1	local (slow)	~8.9GB	how to run ›	+2
33	MOSS-TTS v1.5 (8B) · fp32 (transformers 5) OpenMOSS	63±2	local (slow)	~24GB	how to run ›
34	GLM-TTS · fp32 (official GLM-TTS repo) Z.ai	63±2	local (slow)	~8.9GB	how to run ›
36	MiniMax Speech-02-HD MiniMax	63±3	api $100.00/M chars		platform.minimax.io ↗
37	Fish Audio S2 Pro Fish Audio	63±1	api $15.00/M chars +local	~11GB	docs.fish.audio ↗	+3
38	Eleven Flash v2.5 ElevenLabs	63±2	api $50.00/M chars		elevenlabs.io ↗
39	IndexTTS (v1) · fp32 on CPU (index-tts repo CLI `indextts`) IndexTeam (bilibili)	62±3	local	~1.2GB	how to run ›
40	F5-TTS v1 Base · fp32 SWivid (SJTU X-LANCE)	62±3	local	~1.4GB	how to run ›
41	Inworld TTS-1.5 Max Inworld	62±3*	api $10.00/M chars		fal.ai ↗
42	ZipVoice · fp32 (ZipVoice repo) k2-fsa (Xiaomi)	62±3	local	~0.6GB	how to run ›
44	Qwen3-TTS 12Hz 0.6B Base · bf16 (qwen-tts) Alibaba (Qwen)	62±2	local	~2.5GB	how to run ›
45	Eleven Flash v2 ElevenLabs	62±3*	api $50.00/M chars		elevenlabs.io ↗
46	DMOSpeech 2 · fp32 (DMOSpeech2 repo) Columbia University / NewsBreak (Yinghao Aaron Li, Xilin Jiang et al.)	62±4†	local	~4.3GB	how to run ›
47	Eleven Turbo v2 ElevenLabs	61±3*	api $50.00/M chars		elevenlabs.io ↗
48	Inworld Realtime TTS-2 Inworld	61±2	api $25.00/M chars		replicate.com ↗
49	CosyVoice2-0.5B · fp32 (CosyVoice repo) Alibaba (FunAudioLLM)	61±1	local	~4.1GB	how to run ›	+1
50	E2 TTS Base (F5-TTS reproduction) · fp32 SWivid (SJTU X-LANCE)	60±2	local	~1.3GB	how to run ›
51	MiniMax Speech-02-Turbo MiniMax	60±3*	api $60.00/M chars		platform.minimax.io ↗
52	TTS-1 OpenAI	59±4	api $15.00/M chars		developers.openai.com ↗
53	MaskGCT · fp32 Amphion (CUHK-Shenzhen)	59±1	local	~6.6GB	how to run ›	+4
55	CosyVoice v3 Plus Alibaba Cloud	58±4*	api $26.00/M chars		www.alibabacloud.com ↗
56	Octave TTS (Octave 1) Hume AI	58±4	api $50.00/M chars		www.hume.ai ↗
57	Sonic 3 Cartesia	58±2	api $65.00/M chars		www.together.ai ↗
60	F5-TTS Base (v0) · fp32 SWivid (SJTU X-LANCE)	57±2	local	~1.4GB	how to run ›
61	TTS-1 HD OpenAI	57±7*	api $30.00/M chars		developers.openai.com ↗
62	Deepgram Aura 2 Deepgram	57±3	api $30.00/M chars		deepgram.com ↗
63	ZONOS2 8B · q4_k (zonos2.cpp) Zyphra	57±2†	local	~5.2GB	how to run ›
64	Supertonic 3 · onnx-fp32 (supertonic Python SDK 1.3.x) Supertone	55±10*	local	~0.4GB	how to run ›
66	Higgs Audio V2 · bf16 (transformers >=5.3.0) Boson AI	55±3	local (slow)	~12GB	how to run ›
67	Fasih-TTS-V1 · fp32 (coqui-tts 0.27.5) NightPrince (Yahya Elnawasany)	55±10*	local	~1.9GB	how to run ›
68	CosyVoice-300M · fp32 (CosyVoice repo) Alibaba (FunAudioLLM)	54±3	local	~2.3GB	how to run ›
69	Chatterbox Multilingual V3 · fp32 (chatterbox) Resemble AI	54±10*	local	~3.2GB	how to run ›
70	VibeVoice 1.5B · bf16 (transformers >=5.17.0) Microsoft	53±6	local	~5.4GB	how to run ›
71	Lightning v3.1 Smallest.ai	53±3*	api $17.50/M chars		smallest.ai ↗
73	Kokoro 82M v1.0 · fp32 (kokoro package 0.9.4) hexgrad	52±3	local	~0.3GB	how to run ›
74	LEMAS-TTS (w/ prosody, multilingual_prosody) · official checkpoint (LEMAS-TTS repo, F5-TTS-based) LEMAS Project	52±6*†	local	~1.5GB	how to run ›
76	Spark-TTS 0.5B · fp32 (Spark-TTS repo CLI) SparkAudio	52±3	local	~3.9GB	how to run ›
77	Llasa-8B · bf16 (transformers + xcodec2) HKUST Audio	52±3	local (slow)	~20GB	how to run ›
78	Gradium TTS Gradium	51±3*	api $38.00/M chars		docs.gradium.ai ↗
79	OpenVoice V1 · fp32 (myshell-ai/OpenVoice) MyShell	51±14*	local	~0.3GB	how to run ›
80	Eleven Multilingual v2 ElevenLabs	51±5	api $100.00/M chars		elevenlabs.io ↗
84	LEMAS-TTS (w/o prosody, multilingual_grl) · official checkpoint (LEMAS-TTS repo, F5-TTS-based) LEMAS Project	50±6*†	local	~1.4GB	how to run ›
85	MMS TTS Amharic · fp32 (transformers VitsModel) Meta	50±16*	local	~0.2GB	how to run ›
86	Chatterbox · fp32 (chatterbox-tts 0.1.7) Resemble AI	50±5	local	~3.2GB	how to run ›
87	Parler-TTS Mini v0.1 · fp32 Hugging Face (Parler-TTS)	49±14*	local	~2.6GB	how to run ›
88	Fish Speech 1.2 · bf16 (fish-speech repo) Fish Audio	49±14*	local	~1.2GB	how to run ›
89	SILMA TTS v1 · fp32 checkpoint (silma-tts 1.0.5) SILMA AI	48±10*	local	~2.7GB	how to run ›
90	FireRedTTS (v1) · fp32 (FireRedTTS repo) FireRedTeam (Xiaohongshu)	48±3	local	~3.1GB	how to run ›
92	Orpheus TTS 3B 0.1 (finetune-prod) · gguf Q4_K_M (orpheus-cpp 0.0.3) Canopy Labs	47±5	local	~2.4GB	how to run ›
93	Kyutai TTS 1.6B en/fr · bf16 (moshi) Kyutai	47±5*	local	~4.1GB	how to run ›
95	Amphion VALL-E (LibriLight 6k) · fp32 (Amphion repo recipe) Amphion (CUHK-Shenzhen)	47±14*	local	~1.6GB	how to run ›
96	Piper (en_US lessac medium) · onnx-fp32 (piper-tts 1.8.0) Open Home Foundation (Piper)	47±5*	local	~0.1GB	how to run ›
97	Vevo (Amphion) · fp32 (Amphion repo) Amphion (CUHK-Shenzhen)	46±3	local	~6.8GB	how to run ›
99	Dia2 2B · bf16 (nari-labs/dia2 repo CLI) Nari Labs	45±6*	local (slow)	~8.1GB	how to run ›
100	Amazon Polly Neural Amazon Web Services	45±4*	api $16.00/M chars		aws.amazon.com ↗
102	GPT-SoVITS v2 · fp32 (repo api_v2.py) RVC-Boss	44±4	local	~1.1GB	how to run ›
103	MeloTTS English · fp32 (MeloTTS repo CLI) MyShell	44±5	local	~0.7GB	how to run ›
104	Parler-TTS Mini v1 · fp32 (parler-tts) Hugging Face (Parler-TTS)	44±6*	local	~3.5GB	how to run ›
106	Nabra-82M v0.1 · fp32 (Kokoro KModel, Oddadmix/kokoro fork) oddadmix	42±10*	local	~0.3GB	how to run ›
107	OpenAudio S1 Mini · bf16 (fish-speech official repo CLI) Fish Audio	41±5	local	~3.6GB	how to run ›
108	Fish Speech 1.5 · bf16 (fish-speech repo) Fish Audio	41±3	local	~1.5GB	how to run ›
109	EMA Lightning · fp32 (PyTorch) EMA Lightning (canberkkkkkk)	40±3*	local	~0GB	how to run ›
111	HierSpeech++ v1.1 · fp32 (repo inference.py) Korea University (Sang-Hoon Lee et al.)	39±3	local	~2.1GB	how to run ›
113	Magpie TTS Multilingual 357M · f16-gguf (NVIDIA NeMo-Speech.cpp) NVIDIA	39±6*	local	~0.7GB	how to run ›
114	Zonos v0.1 · bf16 Zyphra	39±4	local	~4.3GB	how to run ›
115	Maya1 · bf16 (transformers + snac) Maya Research	38±6*	local	~6.7GB	how to run ›
116	Gemini 3.8 Flash TTS Google	37±3*	api price unknown		ai.google.dev ↗
117	Pheme 300M · fp32 (PolyAI-LDN/pheme repo) PolyAI	37±3	local	~3.6GB	how to run ›
118	VoiceCraft 830M TTSEnhanced · fp32 (jasonppy/VoiceCraft repo) Puyuan Peng et al. (UT Austin)	37±8*	local	~4.5GB	how to run ›
119	Gemini 3.8 Flash-Lite TTS Google	37±3*	api price unknown		ai.google.dev ↗
120	Eleven v4 ElevenLabs	37±4*	api $80.00/M chars		elevenlabs.io ↗
121	Amphion VALL-E v2 recipe (MLS) · fp32 (Amphion repo) Amphion (CUHK-Shenzhen)	36±3	local	~2.9GB	how to run ›
122	NeuTTS Air · bf16 (neutts 1.4.1) Neuphonic	36±10*	local	~5GB	how to run ›
123	Eleven v4 Turbo ElevenLabs	36±4*	api $40.00/M chars		elevenlabs.io ↗
124	VITS (VCTK multi-speaker) · fp32 (jaywalnut310/vits repo) Kakao Enterprise (Jaehyeon Kim et al.)	34±5	local	~0.2GB	how to run ›
125	Sesame CSM-1B Sesame	34±5	api $7.00/M chars		deepinfra.com ↗
126	VoiceCraft 830M (giga830M) · fp32 (VoiceCraft repo) Puyuan Peng et al. (UT Austin)	32±4†	local	~4.5GB	how to run ›
127	VALL-E X (Plachtaa open implementation) · fp32 (Plachtaa/VALL-E-X repo) Plachtaa (community)	31±4	local	~1.6GB	how to run ›
128	OuteTTS 0.2 500M · fp32 (outetts 0.3.3) OuteAI	31±6*	local	~1.4GB	how to run ›
130	StyleTTS 2 (LibriTTS) · fp32 (styletts2 PyPI package 0.1.6) Yinghao Aaron Li et al. (Columbia University)	30±2	local	~0.9GB	how to run ›	+2
131	XTTS v2 · fp32 (coqui-tts 0.27.5) Coqui	30±2	local	~1.9GB	how to run ›	+7
132	Tortoise TTS v2 · fp16 (neonbjb/tortoise-tts) Tortoise	29±2	local	~5.6GB	how to run ›	+2
133	Parler-TTS Large v1 · fp32 Hugging Face (Parler-TTS)	29±3	local (slow)	~9.3GB	how to run ›
134	WhisperSpeech v1.95 Medium (7-language) · fp16 (whisperspeech 0.8.9) Collabora	28±3	local	~3.2GB	how to run ›
136	HierSpeech++ (LibriTTS-960) · fp32 (repo inference.py) Korea University (Sang-Hoon Lee et al.)	28±5	local	~2.1GB	how to run ›
137	Piper (tr_TR dfki medium) · onnx-fp32 (piper-tts) Open Home Foundation (Piper)	26±4*	local	~0.1GB	how to run ›
138	OpenVoice V2 · fp32 (myshell-ai/OpenVoice) MyShell	26±3	local	~0.8GB	how to run ›
139	Bark · fp32 (transformers) Suno	25±3	local	~4.5GB	how to run ›
140	SpeechT5 TTS · fp32 (transformers) Microsoft	23±3	local	~0.7GB	how to run ›
141	MetaVoice 1B v0.1 · bf16 (metavoiceio/metavoice-src) MetaVoice	20±3	local (slow)	~9.7GB	how to run ›
143	YourTTS · fp32 (coqui-tts) Coqui / Edresson Casanova et al.	18±3	local	~0.4GB	how to run ›
144	MMS TTS Turkish · fp32 (transformers VitsModel) Meta	18±4*	local	~0.2GB	how to run ›
146	Amphion NaturalSpeech 2 (LibriTTS) · fp32 (Amphion repo recipe) Amphion (CUHK-Shenzhen)	14±3	local	~1.7GB	how to run ›
147	Glow-TTS Common Voice (Turkish) · fp32 · CPU Coqui TTS / Fatih Akademi	11±4*	local	~4GB	how to run ›
—	Azure HD 2.5 Microsoft	no reviewed results	api $22.00/M chars		azure.microsoft.com ↗
—	Breeze TTS 2 BreezeBlue	no reviewed results	api $28.00/M chars +local	~8.8GB	breezeblue.ai ↗
—	Chatterbox Turbo · fp32 (chatterbox-tts 0.1.7) Resemble AI	no reviewed results	local	~3GB	how to run ›
—	Dia 1.6B (0626) · fp32 (nari-labs/dia repo) Nari Labs	no reviewed results	local	~6.8GB	how to run ›
—	Eleven v3 Conversational ElevenLabs	no reviewed results	api $40.00/M chars		elevenlabs.io ↗
—	FastPitch (Coqui LJ Speech) · fp32 · CPU Coqui TTS / Eren Gölge @erogol	no reviewed results	local	~4GB	how to run ›
—	FastPitch 1.1 + WaveGlow · fp32 (NVIDIA DeepLearningExamples FastPitch, NGC container) NVIDIA	not scored yet	local	~1.2GB	how to run ›
—	FastSpeech 2 + HiFi-GAN (SpeechBrain LJ Speech) · fp32 · CPU SpeechBrain	no reviewed results	local	~4GB	how to run ›
—	Glow-TTS (Coqui LJ Speech) · fp32 · CPU Coqui TTS / Eren Gölge @erogol	no reviewed results	local	~4GB	how to run ›
—	Glow-TTS Mai (Ukrainian) · fp32 · CPU Coqui TTS / @robinhad	no reviewed results	local	~4GB	how to run ›
—	Grad-TTS + HiFi-GAN (10 steps) · fp32 (huawei-noah/Speech-Backbones Grad-TTS repo) Grad-TTS authors / Huawei Noah	not yet calibrated	local	~0.1GB	how to run ›
—	Kitten TTS Mini 0.8 · onnx (kittenml 0.9.3) KittenML	no reviewed results	local	~0.1GB	how to run ›
—	KittenTTS 2 · ternary-packed LM (kittenml 0.9.3) KittenML	no reviewed results	local	~2.1GB	how to run ›
—	Lightning v3.1 Pro Smallest.ai	no reviewed results	api $19.50/M chars		smallest.ai ↗
—	LightSpeech + Parallel WaveGAN · fp32 (microsoft/NeuralSpeech LightSpeech repo, user text) Microsoft NeuralSpeech	no reviewed results	local	~0GB	how to run ›
—	MiniMax Speech 2.6 HD MiniMax	no reviewed results	api $100.00/M chars		platform.minimax.io ↗
—	Neural HMM TTS (Coqui LJ Speech) · fp32 · CPU Coqui TTS / Shivam Metha @shivammehta25	no reviewed results	local	~4GB	how to run ›
—	Octave 2 Hume AI	no reviewed results	api $25.00/M chars		www.hume.ai ↗
—	OverFlow (Coqui LJ Speech) · fp32 · CPU Coqui TTS / Eren Gölge @erogol	no reviewed results	local	~4GB	how to run ›
—	Pocket TTS (English) · fp32 (pocket-tts 3.3.0) Kyutai	no reviewed results	local	~0.2GB	how to run ›
—	Polly Generative Amazon	no reviewed results	api $30.00/M chars		aws.amazon.com ↗
—	PriorGrad acoustic + HiFi-GAN · fp32 (microsoft/NeuralSpeech PriorGrad-acoustic repo, T=12 fast sampling) Microsoft NeuralSpeech	no reviewed results	local	~0.3GB	how to run ›
—	Sonic 3.6 Cartesia	no reviewed results	api $38.00/M chars		cartesia.ai ↗
—	SpeedySpeech (Coqui LJ Speech) · fp32 · CPU Coqui TTS / Eren Gölge @erogol	no reviewed results	local	~4GB	how to run ›
—	Tacotron 2 + WaveGlow (NVIDIA) · fp32 (torch.hub NVIDIA/DeepLearningExamples:torchhub) NVIDIA	not scored yet	local	~1.2GB	how to run ›
—	Tacotron 2 DDC (Coqui LJ Speech) · fp32 · CPU Coqui TTS / Eren Gölge @erogol	no reviewed results	local	~4GB	how to run ›
—	VITS (LJ Speech) · fp32 (jaywalnut310/vits repo) Kakao / KAIST	not yet calibrated	local	~0.2GB	how to run ›
—	VITS Common Voice (Portuguese) · fp32 · CPU Coqui TTS / @NeonGeckoCom	no reviewed results	local	~4GB	how to run ›
—	VITS CSS10 (French) · fp32 · CPU Coqui TTS / @NeonGeckoCom	no reviewed results	local	~4GB	how to run ›
—	VITS CSS10 (Spanish) · fp32 · CPU Coqui TTS / @NeonGeckoCom	no reviewed results	local	~4GB	how to run ›
—	VITS Thorsten (German) · fp32 · CPU Coqui TTS / @thorstenMueller	no reviewed results	local	~4GB	how to run ›

+ 38 more don't fit your machine or budget · show them

187 of 187 catalogue models listed here · sources and licences

* new or thinly tested: ranked on too few benchmark results · † only the maker’s own numbers so far · ~RAM is estimated; full context may need more memory