Controlled Voice Arena LeaderboardArtificial Analysis

Compare Text to Speech (TTS) models using the same 8 cloned voices (4 US, 4 UK).

Range
Model
Elo
Samples
Released
API Pricing 1
11-2
Alibaba logo
Qwen-Audio-3.1-TTS-Plus
1184±1695% interval 1168 to 1200
1,588Sept 2026$19.3 /1M chars
21-3
ElevenLabs logo
Eleven v4 Turbo
1169±1595% interval 1154 to 1184
1,836Sept 2026$40.0 /1M chars
32-3
ElevenLabs logo
Eleven v4
1159±1595% interval 1144 to 1174
1,786Sept 2026$80.0 /1M chars
44-5
Inworld logo
Realtime TTS-2
1143±1595% interval 1128 to 1158
2,444Aug 2026$20.8 /1M chars
54-6
Cartesia logo
Sonic 3.6
1138±1595% interval 1123 to 1153
2,466Aug 2026$49.0 /1M chars
65-7
Alibaba logo
Qwen-Audio-3.0-TTS-Plus
1126±1595% interval 1111 to 1141
2,211Sept 2026$19.3 /1M chars
76-7
Cartesia logo
Sonic 3.5
1115±1395% interval 1102 to 1128
4,439May 2026$49.0 /1M chars
88-10
StepFun logo
StepAudio 2.5 TTS
1071±1495% interval 1057 to 1085
3,124Apr 2026$85.0 /1M chars
98-10
ElevenLabs logo
Eleven v3
1071±1395% interval 1058 to 1084
3,779Feb 2026$100.0 /1M chars
108-13
Inworld logo
Realtime TTS-2 Flash
1067±1495% interval 1053 to 1081
2,567Aug 2026$10.4 /1M chars
119-14
Google logo
Gemini 3.8 Flash-Lite TTS
1057±1495% interval 1043 to 1071
2,519Sept 2026$11.0 /1M chars
1210-14
ElevenLabs logo
v3 Conversational
1056±1495% interval 1042 to 1070
2,214Aug 2026$50.0 /1M chars
1310-14
SpaceXAI logo
SpaceXAI TTS
1054±1495% interval 1040 to 1068
2,847Mar 2026$15.0 /1M chars
1411-15
Google logo
Gemini 3.8 Flash TTS
1048±1495% interval 1034 to 1062
2,478Sept 2026$16.5 /1M chars
1514-15
SpeechifyAI logo
Simba 3.0
1035±1495% interval 1021 to 1049
2,710Feb 2026$6.6 /1M chars
1616-21
BreezeBlue logo
Breeze TTS 2Hugging Face
1019±1595% interval 1004 to 1034
2,194Aug 2026$34.0 /1M chars
1716-21
MiniMax logo
Speech 2.8 HD
1018±1395% interval 1005 to 1031
3,336Feb 2026$100.0 /1M chars
1816-23
Gradium logo
Gradium TTS (Aug 2026)
1016±1595% interval 1001 to 1031
2,213Aug 2026$47.2 /1M chars
1916-22
MiniMax logo
Speech 2.8 Turbo
1015±1395% interval 1002 to 1028
3,548Feb 2026$60.0 /1M chars
2016-25
Fish Audio logo
Fish Audio S2.1 Pro
1012±1495% interval 998 to 1026
2,597Jun 2026$15.0 /1M chars
2116-25
Fish Audio logo
OpenAudio S1 MiniHugging Face
1009±1495% interval 995 to 1023
2,608Jun 2025$15.0 /1M chars
2218-25
Fish Audio logo
Fish Audio S2 ProHugging Face
1003±1395% interval 990 to 1016
4,008Mar 2026$15.0 /1M chars
2320-26
Mistral logo
Voxtral TTSHugging Face
1001±1495% interval 987 to 1015
2,593Mar 2026$16.0 /1M chars
2424
Fish Audio logo
OpenAudio S1
1000AnchorFixed at 1000 as the reference point for this board's Elo scale.
5,621Jun 2025$15.0 /1M chars
2520-27
ElevenLabs logo
Multilingual v2
999±1495% interval 985 to 1013
3,401Aug 2023$100.0 /1M chars
2623-27
Fish Audio logo
Fish Audio S1
988±1495% interval 974 to 1002
3,427Oct 2025$15.0 /1M chars
2724-27
ElevenLabs logo
Flash v2.5
987±1495% interval 973 to 1001
3,392Dec 2024$50.0 /1M chars
2828-29
ElevenLabs logo
Turbo v2.5
972±1495% interval 958 to 986
2,658Jul 2024$50.0 /1M chars
2928-30
KugelAudio logo
Kugel 3
971±1495% interval 957 to 985
3,912Jun 2026N/A
3028-32
Boson AI logo
Higgs Audio V3 TTSHugging Face
958±1595% interval 943 to 973
2,310Jun 2026$15.0 /1M chars
3130-34
Cartesia logo
Sonic 3
946±1595% interval 931 to 961
2,692Oct 2025$49.0 /1M chars
3230-35
Soniox logo
Soniox TTS Real-Time v2
944±1595% interval 929 to 959
2,146Aug 2026$14.2 /1M chars
3331-36
NVIDIA logo
Magpie TTS Zeroshot
937±1795% interval 920 to 954
1,051Jul 2026N/A
3431-36
Krafton logo
Raon SpeechLM
932±1595% interval 917 to 947
2,153Apr 2026N/A
3532-36
Resemble AI logo
ChatterboxGitHub
931±1595% interval 916 to 946
2,664May 2025$25.0 /1M chars
3634-36
Noiz logo
Noiz TTS
921±1595% interval 906 to 936
2,214Nov 2025N/A
3737
Smallest.ai logo
Lightning v3.1
887±1695% interval 871 to 903
2,296Mar 2026$25.0 /1M chars
3838
Coqui logo
XTTS v2Hugging Face
836±1795% interval 819 to 853
2,460Nov 2023$40.4 /1M chars
3939
OpenVoice logo
OpenVoice v2Hugging Face
806±1895% interval 788 to 824
2,169Apr 2024$8.3 /1M chars

1 API Pricing reflects the cost to generate 1M characters on the model creator's API with the model's default settings

2 Open Weights includes models that can be self-hosted using downloadable model weights or a provider-supplied inference runtime. Open Weights Commercial includes models whose published license permits commercial use, including use subject to license conditions (e.g. an NVIDIA AI Enterprise subscription), and excludes models that require a separately negotiated commercial agreement.