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 | Pricing 1 | |
|---|---|---|---|---|---|---|
| 1 | 1-2 | HeyGen Voice | 1188±1395% interval 1175 to 1201 | 2,356 | Oct 2026 | $30.00 /1M chars |
| 2 | 1-3 | Qwen-Audio-3.1-TTS-Plus | 1180±1395% interval 1167 to 1193 | 2,691 | Sept 2026 | $19.31 /1M chars |
| 3 | 2-3 | HeyGen Voice Turbo | 1172±1695% interval 1156 to 1188 | 2,172 | Oct 2026 | $4.00 /1M chars |
| 4 | 4-7 | Eleven v4 Turbo | 1153±1295% interval 1141 to 1165 | 3,071 | Sept 2026 | $40.00 /1M chars |
| 5 | 4-7 | Eleven v4 | 1153±1295% interval 1141 to 1165 | 3,008 | Sept 2026 | $80.00 /1M chars |
| 6 | 4-7 | Sonic 3.6 | 1147±1295% interval 1135 to 1159 | 3,798 | Aug 2026 | $49.00 /1M chars |
| 7 | 4-7 | Realtime TTS-2 | 1145±1295% interval 1133 to 1157 | 3,610 | Aug 2026 | $20.83 /1M chars |
| 8 | 8 | Qwen-Audio-3.0-TTS-Plus | 1128±1295% interval 1116 to 1140 | 3,438 | Sept 2026 | $19.31 /1M chars |
| 9 | 9 | Sonic 3.5 | 1116±1195% interval 1105 to 1127 | 5,911 | May 2026 | $49.00 /1M chars |
| 10 | 10-11 | Lightning V3.1 Pro (Oct 2026) | 1100±1595% interval 1085 to 1115 | 2,120 | Oct 2026 | $19.50 /1M chars |
| 11 | 10-12 | AssemblyAI-TTS-1 | 1091±1595% interval 1076 to 1106 | 2,017 | Oct 2026 | $15.00 /1M chars |
| 12 | 11-15 | Simba 3.2 | 1083±1295% interval 1071 to 1095 | 3,500 | Jul 2026 | $6.58 /1M chars |
| 13 | 12-21 | 1075±1595% interval 1060 to 1090 | 2,121 | Oct 2026 | $15.00 /1M chars | |
| 14 | 12-19 | Eleven v3 | 1074±1195% interval 1063 to 1085 | 5,014 | Feb 2026 | $100.00 /1M chars |
| 15 | 12-19 | StepAudio 2.5 TTS | 1074±1195% interval 1063 to 1085 | 4,397 | Apr 2026 | $85.00 /1M chars |
| 16 | 12-22 | Voxtral 3 TTS | 1069±1595% interval 1054 to 1084 | 2,290 | Oct 2026 | $13.00 /1M chars |
| 17 | 13-22 | 1067±1495% interval 1053 to 1081 | 2,429 | Oct 2026 | N/A | |
| 18 | 13-21 | Realtime TTS-2 Flash | 1066±1295% interval 1054 to 1078 | 3,758 | Aug 2026 | $10.42 /1M chars |
| 19 | 13-22 | v3 Conversational | 1065±1295% interval 1053 to 1077 | 3,268 | Aug 2026 | $50.00 /1M chars |
| 20 | 14-24 | 1062±1395% interval 1049 to 1075 | 2,670 | Oct 2026 | $15.00 /1M chars | |
| 21 | 16-24 | SpaceXAI TTS | 1061±1295% interval 1049 to 1073 | 3,950 | Mar 2026 | $15.00 /1M chars |
| 22 | 19-24 | Gradium TTS (Sep 2026) | 1054±1295% interval 1042 to 1066 | 2,825 | Sept 2026 | $47.78 /1M chars |
| 23 | 20-24 | Gemini 3.8 Flash TTS | 1052±1295% interval 1040 to 1064 | 3,479 | Sept 2026 | $16.49 /1M chars |
| 24 | 20-25 | Gemini 3.8 Flash-Lite TTS | 1052±1295% interval 1040 to 1064 | 3,596 | Sept 2026 | $11.03 /1M chars |
| 25 | 24-25 | Simba 3.0 | 1040±1295% interval 1028 to 1052 | 3,695 | Feb 2026 | $6.58 /1M chars |
| 26 | 26-29 | Speech 2.8 HD | 1024±1295% interval 1012 to 1036 | 4,175 | Feb 2026 | $100.00 /1M chars |
| 27 | 26-29 | Speech 2.8 Turbo | 1023±1195% interval 1012 to 1034 | 4,535 | Feb 2026 | $60.00 /1M chars |
| 28 | 26-32 | 1020±1295% interval 1008 to 1032 | 3,043 | Aug 2026 | $34.00 /1M chars | |
| 29 | 26-32 | Gradium TTS (Aug 2026) | 1019±1295% interval 1007 to 1031 | 2,986 | Aug 2026 | $47.22 /1M chars |
| 30 | 28-34 | 1011±1195% interval 1000 to 1022 | 5,033 | Mar 2026 | $15.00 /1M chars | |
| 31 | 27-35 | Fish Audio S2.1 Pro (Oct 2026) | 1010±1395% interval 997 to 1023 | 2,364 | Oct 2026 | $15.00 /1M chars |
| 32 | 28-35 | 1008±1295% interval 996 to 1020 | 3,290 | Mar 2026 | $16.00 /1M chars | |
| 33 | 29-35 | 1006±1395% interval 993 to 1019 | 3,180 | Jun 2025 | $15.00 /1M chars | |
| 34 | 30-35 | Multilingual v2 | 1002±1295% interval 990 to 1014 | 4,169 | Aug 2023 | $100.00 /1M chars |
| 35 | 35 | OpenAudio S1 | 1000AnchorFixed at 1000 as the reference point for this board's Elo scale. | 7,053 | Jun 2025 | $15.00 /1M chars |
| 36 | 35-38 | Flash v2.5 | 989±1295% interval 977 to 1001 | 4,171 | Dec 2024 | $50.00 /1M chars |
| 37 | 36-38 | Fish Audio S1 | 987±1295% interval 975 to 999 | 4,007 | Oct 2025 | $15.00 /1M chars |
| 38 | 36-38 | Turbo v2.5 | 983±1395% interval 970 to 996 | 3,057 | Jul 2024 | $50.00 /1M chars |
| 39 | 39-42 | Kugel 3 | 969±1295% interval 957 to 981 | 4,395 | Jun 2026 | N/A |
| 40 | 39-42 | 965±1395% interval 952 to 978 | 2,833 | Jun 2026 | $15.00 /1M chars | |
| 41 | 39-42 | Soniox TTS Real-Time v2 | 963±1395% interval 950 to 976 | 2,566 | Aug 2026 | $14.23 /1M chars |
| 42 | 39-42 | Sonic 3 | 959±1395% interval 946 to 972 | 2,993 | Oct 2025 | $49.00 /1M chars |
| 43 | 42-46 | 943±1795% interval 926 to 960 | 1,060 | Jul 2026 | N/A | |
| 44 | 43-45 | 941±1395% interval 928 to 954 | 2,938 | May 2025 | $25.00 /1M chars | |
| 45 | 43-46 | Raon SpeechLM | 933±1395% interval 920 to 946 | 2,513 | Apr 2026 | N/A |
| 46 | 44-46 | Noiz TTS | 928±1495% interval 914 to 942 | 2,556 | Nov 2025 | N/A |
| 47 | 47 | Lightning v3.1 | 889±1495% interval 875 to 903 | 2,622 | Mar 2026 | $25.00 /1M chars |
| 48 | 48 | 850±1595% interval 835 to 865 | 2,736 | Nov 2023 | $40.44 /1M chars | |
| 49 | 49 | 825±1695% interval 809 to 841 | 2,351 | Apr 2024 | $8.33 /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.