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 | 1-2 |  Eleven v4 Turbo | 1336±1795% interval 1319 to 1353 | 2,085 | Sept 2026 | $40.0 /1M chars | |
| 2 | 1-2 | Eleven v4 | 1332±1695% interval 1316 to 1348 | 2,384 | Sept 2026 | $80.0 /1M chars | |
| 3 | 3-4 |  Qwen-Audio-3.1-TTS-Plus | 1307±1695% interval 1291 to 1323 | 2,030 | Sept 2026 | $19.3 /1M chars | |
| 4 | 4-5 |  Sonic 3.6 | 1292±1495% interval 1278 to 1306 | 2,842 | Aug 2026 | $49.0 /1M chars | |
| 5 | 4-6 |  Gemini 3.8 Flash TTS | 1280±1495% interval 1266 to 1294 | 3,091 | Sept 2026 | $16.5 /1M chars | |
| 6 | 5-7 |  Realtime TTS-2 | 1270±1495% interval 1256 to 1284 | 2,254 | Aug 2026 | $20.8 /1M chars | |
| 7 | 6-7 | Qwen-Audio-3.0-TTS-Plus | 1266±1495% interval 1252 to 1280 | 2,507 | Sept 2026 | $19.3 /1M chars | |
| 8 | 8-10 | Gemini 3.8 Flash-Lite TTS | 1252±1395% interval 1239 to 1265 | 3,078 | Sept 2026 | $11.0 /1M chars | |
| 9 | 8-10 |  Fish Audio Drama 3 Preview  | 1248±1595% interval 1233 to 1263 | 2,237 | Oct 2026 | $15.0 /1M chars | |
| 10 | 8-10 |  Simba 3.2 | 1246±1295% interval 1234 to 1258 | 3,436 | Jul 2026 | $6.6 /1M chars | |
| 11 | 11-13 |  Breeze TTS 2 | 1230±1495% interval 1216 to 1244 | 2,378 | Aug 2026 | $34.0 /1M chars | |
| 12 | 11-17 |  Voxtral 3 TTS | 1219±1595% interval 1204 to 1234 | 1,710 | Oct 2026 | $13.0 /1M chars | |
| 13 | 11-17 | Realtime TTS-2 Flash | 1219±1395% interval 1206 to 1232 | 2,667 | Aug 2026 | $10.4 /1M chars | |
| 14 | 12-17 | v3 Conversational | 1215±1395% interval 1202 to 1228 | 2,837 | Aug 2026 | $50.0 /1M chars | |
| 15 | 12-17 |  StepAudio 2.5 TTS (Aug 2026) | 1214±1495% interval 1200 to 1228 | 2,254 | Aug 2026 | $85.0 /1M chars | |
| 16 | 12-17 | Gemini 3.1 Flash TTS | 1213±1195% interval 1202 to 1224 | 4,327 | Apr 2026 | $18.3 /1M chars | |
| 17 | 12-17 |  Luna TTS | 1210±1295% interval 1198 to 1222 | 3,598 | Jun 2026 | $15.0 /1M chars | |
| 18 | 18-22 | Sonic 3.5 | 1192±1195% interval 1181 to 1203 | 3,881 | May 2026 | $49.0 /1M chars | |
| 19 | 18-22 |  Ming-3.0-tts | 1191±1495% interval 1177 to 1205 | 2,169 | Oct 2026 | N/A | |
| 20 | 18-24 | Fish Audio S2.1 Pro (Oct 2026) | 1186±1495% interval 1172 to 1200 | 2,123 | Oct 2026 | $15.0 /1M chars | |
| 21 | 18-24 | Eleven v3 | 1184±1095% interval 1174 to 1194 | 5,140 | Feb 2026 | $100.0 /1M chars | |
| 22 | 18-25 |  Soniox TTS Real-Time v2 | 1182±1195% interval 1171 to 1193 | 5,436 | Aug 2026 | $14.2 /1M chars | |
| 23 | 20-25 |  Gradium TTS (Sep 2026) | 1177±1395% interval 1164 to 1190 | 2,245 | Sept 2026 | $47.8 /1M chars | |
| 24 | 21-25 |  Speech 2.8 HD | 1176±1095% interval 1166 to 1186 | 5,246 | Feb 2026 | $100.0 /1M chars | |
| 25 | 22-25 |  Lightning V3.1 Pro (Jul 2026) | 1171±1295% interval 1159 to 1183 | 2,859 | Jul 2026 | $19.5 /1M chars | |
| 26 | 26-29 |  Falcon 2 | 1158±1395% interval 1145 to 1171 | 2,501 | Jul 2026 | $10.0 /1M chars | |
| 27 | 26-29 | Speech 2.8 Turbo | 1156±1095% interval 1146 to 1166 | 5,016 | Feb 2026 | $60.0 /1M chars | |
| 28 | 26-31 | Gradium TTS (Aug 2026) | 1154±1395% interval 1141 to 1167 | 2,493 | Aug 2026 | $47.2 /1M chars | |
| 29 | 26-32 |  SpaceXAI TTS | 1152±1395% interval 1139 to 1165 | 1,935 | Mar 2026 | $15.0 /1M chars | |
| 30 | 28-32 |  Azure HD 2.5 | 1143±1195% interval 1132 to 1154 | 3,024 | Mar 2026 | $22.0 /1M chars | |
| 31 | 29-33 |  Async Flash v1.5 | 1141±1295% interval 1129 to 1153 | 2,725 | May 2026 | $10.1 /1M chars | |
| 32 | 30-33 | Step TTS 2 (Mar 2026) | 1140±1295% interval 1128 to 1152 | 2,835 | Feb 2026 | $40.0 /1M chars | |
| 33 | 31-36 | Speech 2.6 HD | 1130±1195% interval 1119 to 1141 | 4,316 | Oct 2025 | $100.0 /1M chars | |
| 34 | 33-36 | Async Pro v1.0 | 1128±1295% interval 1116 to 1140 | 2,602 | May 2026 | $20.2 /1M chars | |
| 35 | 33-36 | Speech 2.6 Turbo | 1122±1195% interval 1111 to 1133 | 4,590 | Oct 2025 | $60.0 /1M chars | |
| 36 | 33-39 | Simba 3.0 | 1119±1195% interval 1108 to 1130 | 3,053 | Feb 2026 | $6.6 /1M chars | |
| 37 | 36-41 | Speech-02-HD | 1111±1195% interval 1100 to 1122 | 4,668 | Mar 2025 | $100.0 /1M chars | |
| 38 | 35-42 | Fish Audio S2 Pro | 1110±1295% interval 1098 to 1122 | 2,750 | Mar 2026 | $15.0 /1M chars | |
| 39 | 35-42 |  Chatterbox HD | 1110±1295% interval 1098 to 1122 | 2,692 | — | May 2025 | $40.0 /1M chars |
| 40 | 37-44 |  TTS-1 HD | 1103±1195% interval 1092 to 1114 | 3,787 | Nov 2023 | $30.0 /1M chars | |
| 41 | 37-47 | Step Audio EditX (Mar 2026) | 1100±1295% interval 1088 to 1112 | 2,359 | Feb 2026 | N/A | |
| 42 | 40-47 | Turbo v2.5 | 1099±1095% interval 1089 to 1109 | 7,058 | Jul 2024 | $50.0 /1M chars | |
| 43 | 40-49 |  Speech 2.6 | 1097±1295% interval 1085 to 1109 | 2,493 | Jul 2026 | N/A | |
| 44 | 40-49 | Multilingual v2 | 1096±1095% interval 1086 to 1106 | 7,520 | Aug 2023 | $100.0 /1M chars | |
| 45 | 41-49 | Speech-02-Turbo | 1092±1195% interval 1081 to 1103 | 4,002 | Mar 2025 | $60.0 /1M chars | |
| 46 | 41-49 | Voxtral TTS | 1091±1295% interval 1079 to 1103 | 2,460 | Mar 2026 | $16.0 /1M chars | |
| 47 | 42-49 | TTS-1 | 1089±1095% interval 1079 to 1099 | 6,307 | Nov 2023 | $15.0 /1M chars | |
| 48 | 43-51 | Gemini 2.5 Flash Lite TTS | 1087±1195% interval 1076 to 1098 | 3,322 | Dec 2025 | $9.2 /1M chars | |
| 49 | 43-52 | OpenAudio S1 | 1087±1195% interval 1076 to 1098 | 4,658 | Jun 2025 | $15.0 /1M chars | |
| 50 | 48-56 | Flash v2.5 | 1078±1195% interval 1067 to 1089 | 5,849 | Dec 2024 | $50.0 /1M chars | |
| 51 | 48-58 | GPT-Realtime-2 | 1077±1295% interval 1065 to 1089 | 2,324 | May 2026 | $191.6 /1M chars | |
| 52 | 49-58 | Studio | 1076±1195% interval 1065 to 1087 | 4,601 | Feb 2023 | $160.0 /1M chars | |
| 53 | 48-59 | Journey | 1075±1295% interval 1063 to 1087 | 2,368 | Dec 2023 | $160.0 /1M chars | |
| 54 | 50-60 | T2A-01-HD | 1069±1195% interval 1058 to 1080 | 5,088 | Jan 2025 | $50.0 /1M chars | |
| 55 | 50-62 |  Maya 2 Flash | 1068±1395% interval 1055 to 1081 | 2,167 | Jul 2026 | N/A | |
| 56 | 50-62 |  Polly Generative | 1067±1195% interval 1056 to 1078 | 3,851 | May 2024 | $30.0 /1M chars | |
| 57 | 52-65 |  Kokoro 82M v1.0 | 1066±1195% interval 1055 to 1077 | 5,450 | Jan 2025 | $0.7 /1M chars | |
| 58 | 51-65 |  Magpie-Multilingual 357M (Feb 2026) | 1065±1295% interval 1053 to 1077 | 2,315 | Feb 2026 | N/A | |
| 59 | 51-65 | Gemini 2.5 Flash TTS (Dec 2025) | 1065±1295% interval 1053 to 1077 | 2,967 | Dec 2025 | $9.2 /1M chars | |
| 60 | 54-65 | Chirp 3: HD | 1062±1295% interval 1050 to 1074 | 3,459 | Mar 2025 | $30.0 /1M chars | |
| 61 | 54-68 |  Octave 2 | 1058±1295% interval 1046 to 1070 | 3,213 | Oct 2025 | $87.5 /1M chars | |
| 62 | 54-68 | Sonic 3 | 1058±1295% interval 1046 to 1070 | 3,580 | Oct 2025 | $49.0 /1M chars | |
| 63 | 56-70 |  MiMo-V2.5-TTS | 1055±1295% interval 1043 to 1067 | 2,141 | - | N/A | |
| 64 | 56-70 |  Flux TTS | 1055±1395% interval 1042 to 1068 | 2,226 | Aug 2026 | $45.0 /1M chars | |
| 65 | 57-70 |  Coda | 1055±1295% interval 1043 to 1067 | 2,485 | May 2026 | $50.0 /1M chars | |
| 66 | 60-72 | Maya 2 Global | 1050±1395% interval 1037 to 1063 | 2,021 | Jul 2026 | N/A | |
| 67 | 61-72 | Async Flash v1.0 | 1050±1195% interval 1039 to 1061 | 4,106 | Jul 2025 | $10.1 /1M chars | |
| 68 | 61-72 | Maya1 | 1048±1295% interval 1036 to 1060 | 3,336 | Jun 2025 | N/A | |
| 69 | 61-74 |  Higgs Audio V3 TTS | 1045±1395% interval 1032 to 1058 | 2,234 | Jun 2026 | $15.0 /1M chars | |
| 70 | 60-78 | OpenAudio S1 Mini | 1045±2095% interval 1025 to 1065 | 1,682 | Jun 2025 | $15.0 /1M chars | |
| 71 | 66-77 | Gemini 2.5 Pro (Dec 2025) | 1040±1295% interval 1028 to 1052 | 2,490 | Dec 2025 | $18.3 /1M chars | |
| 72 | 67-77 | Octave TTS | 1039±1195% interval 1028 to 1050 | 4,741 | Feb 2025 | $87.5 /1M chars | |
| 73 | 69-78 |  Bland Speech v3 | 1035±1295% interval 1023 to 1047 | 2,542 | Aug 2026 | $40.0 /1M chars | |
| 74 | 66-79 | Azure Neural | 1034±2095% interval 1014 to 1054 | 1,156 | Sept 2018 | $15.0 /1M chars | |
| 75 | 71-79 | T2A-01-Turbo | 1031±1195% interval 1020 to 1042 | 5,000 | Jan 2025 | $30.0 /1M chars | |
| 76 | 71-79 | Polly Long-Form | 1031±1395% interval 1018 to 1044 | 1,986 | Nov 2023 | $100.0 /1M chars | |
| 77 | 71-79 | MAI-Voice-1 | 1030±1295% interval 1018 to 1042 | 2,363 | Apr 2026 | N/A | |
| 78 | 73-79 | Chatterbox  | 1025±1195% interval 1014 to 1036 | 4,724 | May 2025 | $25.0 /1M chars | |
| 79 | 74-79 | Lightning v3.1 | 1023±1295% interval 1011 to 1035 | 2,356 | Mar 2026 | $25.0 /1M chars | |
| 80 | 80-81 |  Raon SpeechLM | 1006±1395% interval 993 to 1019 | 2,216 | Apr 2026 | N/A | |
| 81 | 81 |  Zonos-v0.1 | 1000AnchorFixed at 1000 as the reference point for this board's Elo scale. | 4,875 | Feb 2025 | $20.0 /1M chars | |
| 82 | 82 | Murf Speech Gen 2 | 988±1295% interval 976 to 1000 | 5,035 | Jul 2024 | $100.0 /1M chars | |
| 83 | 83-85 | VibeVoice 1.5B | 956±1495% interval 942 to 970 | 2,160 | Aug 2025 | N/A | |
| 84 | 83-86 |  OpenVoice v2 | 955±1395% interval 942 to 968 | 3,085 | — | Apr 2024 | $8.3 /1M chars |
| 85 | 83-87 |  Neuphonic TTS | 943±1495% interval 929 to 957 | 2,487 | Oct 2024 | $20.8 /1M chars | |
| 86 | 84-87 | Qwen3 TTS Flash | 942±1495% interval 928 to 956 | 2,220 | Sept 2025 | $11.0 /1M chars | |
| 87 | 86-89 | Qwen3 TTS | 930±1395% interval 917 to 943 | 3,113 | Jan 2026 | $13.8 /1M chars | |
| 88 | 87-90 |  Kugel 3 | 926±1395% interval 913 to 939 | 4,394 | Jun 2026 | N/A | |
| 89 | 87-90 | WaveNet | 919±1295% interval 907 to 931 | 6,399 | Sept 2016 | $4.0 /1M chars | |
| 90 | 88-91 |  XTTS v2 | 915±1595% interval 900 to 930 | 2,703 | — | Nov 2023 | $40.4 /1M chars |
| 91 | 89-94 | Mist V2 | 905±1595% interval 890 to 920 | 3,046 | Mar 2025 | $31.1 /1M chars | |
| 92 | 91-95 | Polly Neural | 897±1495% interval 883 to 911 | 2,972 | Jul 2019 | $16.0 /1M chars | |
| 93 | 91-95 | Neural2 | 895±1295% interval 883 to 907 | 6,683 | Jun 2022 | $16.0 /1M chars | |
| 94 | 91-95 |  StyleTTS 2 | 890±1595% interval 875 to 905 | 2,675 | — | Jun 2023 | $2.8 /1M chars |
| 95 | 92-95 | Standard | 888±1295% interval 876 to 900 | 6,765 | Aug 2018 | $4.0 /1M chars | |
| 96 | 96 |  Noiz TTS | 873±1595% interval 858 to 888 | 2,426 | Nov 2025 | N/A | |
| 97 | 97-98 |  MetaVoice v1 | 847±1895% interval 829 to 865 | 1,654 | — | Feb 2024 | N/A |
| 98 | 97-98 | Polly Standard | 831±1695% interval 815 to 847 | 3,581 | 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 Turbo (Elo 1336), 2. Eleven v4 (Elo 1332), 3. Qwen-Audio-3.1-TTS-Plus (Elo 1307), 4. Sonic 3.6 (Elo 1292), and 5. Gemini 3.8 Flash TTS (Elo 1280). 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 1066. Other affordable options include StyleTTS 2 at $2.82 per 1M characters.
The top 5 open weights Text to Speech models are: 1. Fish Audio Drama 3 Preview (Elo 1248), 2. Breeze TTS 2 (Elo 1230), 3. Ming-3.0-tts (Elo 1191), 4. Fish Audio S2 Pro (Elo 1110), and 5. Step Audio EditX (Mar 2026) (Elo 1100).