Provider Voice Arena LeaderboardArtificial Analysis
Compare Text to Speech (TTS) models using each provider's own native voices.
Range | Model | Elo | Samples | Arena Voices | Released | API Pricing 1 | |
|---|---|---|---|---|---|---|---|
| 1 | 1 |  Eleven v4 | 1321±1895% interval 1303 to 1339 | 1,930 | Sept 2026 | $80.0 /1M chars | |
| 2 | 2-4 |  Qwen-Audio-3.1-TTS-Plus | 1292±1895% interval 1274 to 1310 | 1,481 | Sept 2026 | $19.3 /1M chars | |
| 3 | 2-4 |  Sonic 3.6 | 1278±1695% interval 1262 to 1294 | 2,220 | Aug 2026 | $49.0 /1M chars | |
| 4 | 3-5 |  Gemini 3.8 Flash TTS | 1275±1695% interval 1259 to 1291 | 2,446 | Sept 2026 | $16.5 /1M chars | |
| 5 | 4-6 | Qwen-Audio-3.0-TTS-Plus | 1261±1595% interval 1246 to 1276 | 1,835 | Sept 2026 | $19.3 /1M chars | |
| 6 | 5-8 |  Realtime TTS-2 | 1251±1695% interval 1235 to 1267 | 1,613 | Aug 2026 | $20.8 /1M chars | |
| 7 | 6-8 |  Simba 3.2 | 1242±1395% interval 1229 to 1255 | 2,732 | Jul 2026 | $6.6 /1M chars | |
| 8 | 6-9 | Gemini 3.8 Flash-Lite TTS | 1242±1595% interval 1227 to 1257 | 2,372 | Sept 2026 | $11.0 /1M chars | |
| 9 | 9-10 |  Luna TTS | 1227±1395% interval 1214 to 1240 | 2,888 | Jun 2026 | $15.0 /1M chars | |
| 10 | 9-14 |  Breeze TTS 2  | 1216±1595% interval 1201 to 1231 | 1,650 | Aug 2026 | $34.0 /1M chars | |
| 11 | 9-14 | Realtime TTS-2 Flash | 1214±1495% interval 1200 to 1228 | 2,016 | Aug 2026 | $10.4 /1M chars | |
| 12 | 10-14 |  StepAudio 2.5 TTS (Aug 2026) | 1210±1595% interval 1195 to 1225 | 1,571 | Aug 2026 | $85.0 /1M chars | |
| 13 | 10-14 | Gemini 3.1 Flash TTS | 1209±1295% interval 1197 to 1221 | 3,706 | Apr 2026 | $18.3 /1M chars | |
| 14 | 10-15 | v3 Conversational | 1203±1495% interval 1189 to 1217 | 2,122 | Aug 2026 | $50.0 /1M chars | |
| 15 | 15-16 | Sonic 3.5 | 1190±1295% interval 1178 to 1202 | 3,196 | May 2026 | $49.0 /1M chars | |
| 16 | 16-19 |  Soniox TTS Real-Time v2 | 1179±1195% interval 1168 to 1190 | 4,851 | Aug 2026 | $14.2 /1M chars | |
| 17 | 16-19 | Eleven v3 | 1174±1195% interval 1163 to 1185 | 4,563 | Feb 2026 | $100.0 /1M chars | |
| 18 | 16-19 |  Speech 2.8 HD | 1173±1195% interval 1162 to 1184 | 4,638 | Feb 2026 | $100.0 /1M chars | |
| 19 | 16-19 |  Lightning V3.1 Pro (Jul 2026) | 1171±1495% interval 1157 to 1185 | 2,183 | Jul 2026 | $19.5 /1M chars | |
| 20 | 20-23 |  Falcon 2 | 1155±1495% interval 1141 to 1169 | 1,944 | Jul 2026 | $10.0 /1M chars | |
| 21 | 20-22 | Speech 2.8 Turbo | 1153±1195% interval 1142 to 1164 | 4,496 | Feb 2026 | $60.0 /1M chars | |
| 22 | 20-23 |  Gradium TTS (Aug 2026) | 1151±1495% interval 1137 to 1165 | 1,864 | Aug 2026 | $47.2 /1M chars | |
| 23 | 21-27 |  Fish Audio S2.1 Pro | 1141±1395% interval 1128 to 1154 | 2,364 | Jun 2026 | $15.0 /1M chars | |
| 24 | 23-30 |  Async Flash v1.5 | 1136±1395% interval 1123 to 1149 | 2,259 | May 2026 | $10.1 /1M chars | |
| 25 | 22-30 |  SpaceXAI TTS | 1136±1595% interval 1121 to 1151 | 1,377 | Mar 2026 | $15.0 /1M chars | |
| 26 | 23-30 | Step TTS 2 (Mar 2026) | 1134±1295% interval 1122 to 1146 | 2,367 | Feb 2026 | $40.0 /1M chars | |
| 27 | 23-30 |  Azure HD 2.5 | 1131±1295% interval 1119 to 1143 | 2,569 | Mar 2026 | $22.0 /1M chars | |
| 28 | 24-32 | Speech 2.6 HD | 1128±1195% interval 1117 to 1139 | 3,961 | Oct 2025 | $100.0 /1M chars | |
| 29 | 24-33 | Simba 3.0 | 1125±1295% interval 1113 to 1137 | 2,624 | Feb 2026 | $6.6 /1M chars | |
| 30 | 26-33 | Speech 2.6 Turbo | 1124±1195% interval 1113 to 1135 | 4,272 | Oct 2025 | $60.0 /1M chars | |
| 31 | 27-33 | Async Pro v1.0 | 1119±1395% interval 1106 to 1132 | 2,156 | May 2026 | $20.2 /1M chars | |
| 32 | 28-33 | Fish Audio S2 Pro | 1117±1295% interval 1105 to 1129 | 2,377 | Mar 2026 | $15.0 /1M chars | |
| 33 | 29-34 | Speech-02-HD | 1114±1195% interval 1103 to 1125 | 4,375 | Mar 2025 | $100.0 /1M chars | |
| 34 | 33-39 |  Speech 2.6 | 1103±1395% interval 1090 to 1116 | 2,102 | Jul 2026 | N/A | |
| 35 | 33-39 |  TTS-1 HD | 1103±1295% interval 1091 to 1115 | 3,557 | Nov 2023 | $30.0 /1M chars | |
| 36 | 33-41 |  Chatterbox HD | 1102±1395% interval 1089 to 1115 | 2,406 | — | May 2025 | $40.0 /1M chars |
| 37 | 34-41 | Turbo v2.5 | 1098±1195% interval 1087 to 1109 | 6,823 | Jul 2024 | $50.0 /1M chars | |
| 38 | 34-41 | Multilingual v2 | 1097±1095% interval 1087 to 1107 | 7,259 | Aug 2023 | $100.0 /1M chars | |
| 39 | 34-44 | Step Audio EditX (Mar 2026) | 1095±1395% interval 1082 to 1108 | 2,046 | Feb 2026 | N/A | |
| 40 | 37-45 | TTS-1 | 1090±1195% interval 1079 to 1101 | 6,104 | Nov 2023 | $15.0 /1M chars | |
| 41 | 36-46 | Speech-02-Turbo | 1090±1295% interval 1078 to 1102 | 3,780 | Mar 2025 | $60.0 /1M chars | |
| 42 | 39-49 |  Voxtral TTS | 1083±1395% interval 1070 to 1096 | 2,138 | Mar 2026 | $16.0 /1M chars | |
| 43 | 40-48 | Gemini 2.5 Flash Lite TTS | 1083±1295% interval 1071 to 1095 | 3,122 | Dec 2025 | $9.2 /1M chars | |
| 44 | 40-48 | OpenAudio S1 | 1082±1195% interval 1071 to 1093 | 4,470 | Jun 2025 | $15.0 /1M chars | |
| 45 | 40-49 | Studio | 1079±1195% interval 1068 to 1090 | 4,411 | Feb 2023 | $160.0 /1M chars | |
| 46 | 42-49 | Flash v2.5 | 1078±1195% interval 1067 to 1089 | 5,666 | Dec 2024 | $50.0 /1M chars | |
| 47 | 40-53 | GPT-Realtime-2 | 1077±1395% interval 1064 to 1090 | 2,041 | May 2026 | $191.6 /1M chars | |
| 48 | 42-54 | Journey | 1074±1395% interval 1061 to 1087 | 2,206 | Dec 2023 | $160.0 /1M chars | |
| 49 | 42-56 |  Maya 2 Flash | 1071±1495% interval 1057 to 1085 | 1,889 | Jul 2026 | N/A | |
| 50 | 47-57 | T2A-01-HD | 1066±1195% interval 1055 to 1077 | 4,920 | Jan 2025 | $50.0 /1M chars | |
| 51 | 47-58 |  Magpie-Multilingual 357M (Feb 2026) | 1065±1395% interval 1052 to 1078 | 2,064 | Feb 2026 | N/A | |
| 52 | 48-58 |  Polly Generative | 1065±1295% interval 1053 to 1077 | 3,705 | May 2024 | $30.0 /1M chars | |
| 53 | 48-58 |  Kokoro 82M v1.0 | 1064±1195% interval 1053 to 1075 | 5,283 | Jan 2025 | $0.7 /1M chars | |
| 54 | 48-60 |  Coda | 1062±1395% interval 1049 to 1075 | 2,244 | May 2026 | $50.0 /1M chars | |
| 55 | 49-62 | Sonic 3 | 1060±1295% interval 1048 to 1072 | 3,456 | Oct 2025 | $49.0 /1M chars | |
| 56 | 50-62 | Gemini 2.5 Flash TTS (Dec 2025) | 1058±1295% interval 1046 to 1070 | 2,837 | Dec 2025 | $9.2 /1M chars | |
| 57 | 50-63 | Chirp 3: HD | 1056±1295% interval 1044 to 1068 | 3,311 | Mar 2025 | $30.0 /1M chars | |
| 58 | 50-63 |  MiMo-V2.5-TTS | 1055±1395% interval 1042 to 1068 | 1,942 | - | N/A | |
| 59 | 54-63 | Async Flash v1.0 | 1052±1195% interval 1041 to 1063 | 3,982 | Jul 2025 | $10.1 /1M chars | |
| 60 | 54-63 |  Octave 2 | 1051±1295% interval 1039 to 1063 | 3,068 | Oct 2025 | $87.5 /1M chars | |
| 61 | 55-64 | Maya1 | 1049±1295% interval 1037 to 1061 | 3,230 | Jun 2025 | N/A | |
| 62 | 54-67 | Maya 2 Global | 1049±1495% interval 1035 to 1063 | 1,776 | Jul 2026 | N/A | |
| 63 | 54-72 | OpenAudio S1 Mini | 1044±2095% interval 1024 to 1064 | 1,682 | Jun 2025 | $15.0 /1M chars | |
| 64 | 61-72 |  Higgs Audio V3 TTS | 1038±1395% interval 1025 to 1051 | 2,040 | Jun 2026 | $15.0 /1M chars | |
| 65 | 61-74 |  Bland Speech v3 | 1037±1395% interval 1024 to 1050 | 2,322 | Aug 2026 | $40.0 /1M chars | |
| 66 | 63-71 | Octave TTS | 1036±1195% interval 1025 to 1047 | 4,632 | Feb 2025 | $87.5 /1M chars | |
| 67 | 59-74 | Azure Neural | 1035±1995% interval 1016 to 1054 | 1,158 | Sept 2018 | $15.0 /1M chars | |
| 68 | 63-74 | MAI-Voice-1 | 1034±1395% interval 1021 to 1047 | 2,203 | Apr 2026 | N/A | |
| 69 | 63-74 | Polly Long-Form | 1033±1395% interval 1020 to 1046 | 1,904 | Nov 2023 | $100.0 /1M chars | |
| 70 | 63-74 | Gemini 2.5 Pro (Dec 2025) | 1033±1395% interval 1020 to 1046 | 2,417 | Dec 2025 | $18.3 /1M chars | |
| 71 | 64-74 | T2A-01-Turbo | 1031±1195% interval 1020 to 1042 | 4,928 | Jan 2025 | $30.0 /1M chars | |
| 72 | 67-74 | Chatterbox  | 1025±1195% interval 1014 to 1036 | 4,622 | May 2025 | $25.0 /1M chars | |
| 73 | 65-74 | Lightning v3.1 | 1024±1395% interval 1011 to 1037 | 2,166 | Mar 2026 | $25.0 /1M chars | |
| 74 | 64-74 | MiMo-V2-TTS | 1024±1495% interval 1010 to 1038 | 1,762 | Mar 2026 | N/A | |
| 75 | 75-76 |  Raon SpeechLM | 1003±1495% interval 989 to 1017 | 2,063 | Apr 2026 | N/A | |
| 76 | 76 |  Zonos-v0.1 | 1000AnchorFixed at 1000 as the reference point for this board's Elo scale. | 4,819 | Feb 2025 | $20.0 /1M chars | |
| 77 | 77 | Murf Speech Gen 2 | 984±1295% interval 972 to 996 | 4,973 | Jul 2024 | $100.0 /1M chars | |
| 78 | 78-80 | VibeVoice 1.5B | 957±1495% interval 943 to 971 | 2,107 | Aug 2025 | N/A | |
| 79 | 78-80 |  OpenVoice v2 | 954±1395% interval 941 to 967 | 2,981 | Apr 2024 | $8.3 /1M chars | |
| 80 | 78-83 | Qwen3 TTS Flash | 944±1495% interval 930 to 958 | 2,128 | Sept 2025 | $11.0 /1M chars | |
| 81 | 79-83 |  Neuphonic TTS | 940±1495% interval 926 to 954 | 2,384 | Oct 2024 | $20.8 /1M chars | |
| 82 | 80-83 |  Kugel 3 | 935±1395% interval 922 to 948 | 4,156 | Jun 2026 | N/A | |
| 83 | 80-83 | Qwen3 TTS | 932±1395% interval 919 to 945 | 2,902 | Jan 2026 | $13.8 /1M chars | |
| 84 | 84-86 | WaveNet | 918±1295% interval 906 to 930 | 6,235 | Sept 2016 | $4.0 /1M chars | |
| 85 | 84-86 |  XTTS v2 | 916±1595% interval 901 to 931 | 2,544 | — | Nov 2023 | $40.4 /1M chars |
| 86 | 84-88 | Mist V2 | 909±1595% interval 894 to 924 | 2,843 | Mar 2025 | $31.1 /1M chars | |
| 87 | 86-91 |  StyleTTS 2 | 895±1595% interval 880 to 910 | 2,451 | — | Jun 2023 | $2.8 /1M chars |
| 88 | 87-91 | Neural2 | 895±1295% interval 883 to 907 | 6,456 | Jun 2022 | $16.0 /1M chars | |
| 89 | 87-91 | Polly Neural | 894±1495% interval 880 to 908 | 2,752 | Jul 2019 | $16.0 /1M chars | |
| 90 | 87-91 | Standard | 886±1295% interval 874 to 898 | 6,506 | Aug 2018 | $4.0 /1M chars | |
| 91 | 87-91 |  Noiz TTS | 885±1595% interval 870 to 900 | 2,006 | Nov 2025 | N/A | |
| 92 | 92 |  MetaVoice v1 | 850±1895% interval 832 to 868 | 1,360 | — | Feb 2024 | N/A |
| 93 | 93 | Polly Standard | 830±1695% interval 814 to 846 | 3,304 | Nov 2016 | $4.0 /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.
Frequently Asked Questions
The top Text to Speech models by Elo rating are: 1. Eleven v4 (Elo 1321), 2. Qwen-Audio-3.1-TTS-Plus (Elo 1292), 3. Sonic 3.6 (Elo 1278), 4. Gemini 3.8 Flash TTS (Elo 1275), and 5. Qwen-Audio-3.0-TTS-Plus (Elo 1261). Rankings are based on blind user votes in the Speech Arena.
Kokoro 82M v1.0 is the most affordable at $0.65 per 1M characters with an Elo score of 1064. Other affordable options include StyleTTS 2 at $2.82 per 1M characters.
The top 5 open weights Text to Speech models are: 1. Breeze TTS 2 (Elo 1216), 2. Fish Audio S2 Pro (Elo 1117), 3. Step Audio EditX (Mar 2026) (Elo 1095), 4. Voxtral TTS (Elo 1083), and 5. Magpie-Multilingual 357M (Feb 2026) (Elo 1065).