Ranking da Arena de voz do provedorArtificial Analysis
Compare modelos de texto para fala (TTS) usando as vozes nativas de cada provedor.
Faixa | Modelo | Elo | Amostras | Vozes da Arena | Lançamento | Preço 1 | |
|---|---|---|---|---|---|---|---|
| 1 | 1-2 |  Eleven v4 Turbo | 1336±16Intervalo de 95%: 1320 a 1352 | 2,277 | Sept 2026 | US$ 40,0 /1 mi de caracteres | |
| 2 | 1-2 | Eleven v4 | 1334±15Intervalo de 95%: 1319 a 1349 | 2,553 | Sept 2026 | US$ 80,0 /1 mi de caracteres | |
| 3 | 3-4 |  Qwen-Audio-3.1-TTS-Plus | 1305±15Intervalo de 95%: 1290 a 1320 | 2,281 | Sept 2026 | US$ 19,3 /1 mi de caracteres | |
| 4 | 3-5 |  Sonic 3.6 | 1296±13Intervalo de 95%: 1283 a 1309 | 3,126 | Aug 2026 | US$ 49,0 /1 mi de caracteres | |
| 5 | 4-6 |  Gemini 3.8 Flash TTS | 1283±13Intervalo de 95%: 1270 a 1296 | 3,394 | Sept 2026 | US$ 16,5 /1 mi de caracteres | |
| 6 | 4-9 |  Fish Audio Drama 3  | 1278±18Intervalo de 95%: 1260 a 1296 | 1,512 | Oct 2026 | US$ 15,0 /1 mi de caracteres | |
| 7 | 6-10 | Qwen-Audio-3.0-TTS-Plus | 1269±13Intervalo de 95%: 1256 a 1282 | 2,794 | Sept 2026 | US$ 19,3 /1 mi de caracteres | |
| 8 | 6-10 |  Realtime TTS-2 | 1267±14Intervalo de 95%: 1253 a 1281 | 2,527 | Aug 2026 | US$ 20,8 /1 mi de caracteres | |
| 9 | 6-10 |  Lightning V3.1 Pro (Oct 2026) | 1266±16Intervalo de 95%: 1250 a 1282 | 1,813 | Oct 2026 | US$ 19,5 /1 mi de caracteres | |
| 10 | 7-11 | Gemini 3.8 Flash-Lite TTS | 1259±13Intervalo de 95%: 1246 a 1272 | 3,382 | Sept 2026 | US$ 11,0 /1 mi de caracteres | |
| 11 | 10-12 |  Simba 3.2 | 1249±12Intervalo de 95%: 1237 a 1261 | 3,760 | Jul 2026 | US$ 6,6 /1 mi de caracteres | |
| 12 | 11-14 | Fish Audio Drama 3 Preview | 1240±14Intervalo de 95%: 1226 a 1254 | 2,544 | Oct 2026 | US$ 15,0 /1 mi de caracteres | |
| 13 | 12-16 |  Breeze TTS 2 | 1228±13Intervalo de 95%: 1215 a 1241 | 2,650 | Aug 2026 | US$ 34,0 /1 mi de caracteres | |
| 14 | 12-19 |  AssemblyAI-TTS-1 | 1227±16Intervalo de 95%: 1211 a 1243 | 1,632 | Oct 2026 | US$ 15,0 /1 mi de caracteres | |
| 15 | 13-19 |  Voxtral 3 TTS | 1223±14Intervalo de 95%: 1209 a 1237 | 2,008 | Oct 2026 | US$ 13,0 /1 mi de caracteres | |
| 16 | 13-19 | Realtime TTS-2 Flash | 1217±12Intervalo de 95%: 1205 a 1229 | 2,993 | Aug 2026 | US$ 10,4 /1 mi de caracteres | |
| 17 | 15-20 | v3 Conversational | 1215±12Intervalo de 95%: 1203 a 1227 | 3,145 | Aug 2026 | US$ 50,0 /1 mi de caracteres | |
| 18 | 14-20 |  StepAudio 2.5 TTS (Aug 2026) | 1214±13Intervalo de 95%: 1201 a 1227 | 2,598 | Aug 2026 | US$ 85,0 /1 mi de caracteres | |
| 19 | 15-20 | Gemini 3.1 Flash TTS | 1212±11Intervalo de 95%: 1201 a 1223 | 4,586 | Apr 2026 | US$ 18,3 /1 mi de caracteres | |
| 20 | 17-20 |  Luna TTS | 1205±11Intervalo de 95%: 1194 a 1216 | 3,862 | Jun 2026 | US$ 15,0 /1 mi de caracteres | |
| 21 | 21-24 | Sonic 3.5 | 1194±11Intervalo de 95%: 1183 a 1205 | 4,125 | May 2026 | US$ 49,0 /1 mi de caracteres | |
| 22 | 21-25 | Fish Audio S2.1 Pro (Oct 2026) | 1188±13Intervalo de 95%: 1175 a 1201 | 2,411 | Oct 2026 | US$ 15,0 /1 mi de caracteres | |
| 23 | 21-25 | Eleven v3 | 1188±10Intervalo de 95%: 1178 a 1198 | 5,388 | Feb 2026 | US$ 100,0 /1 mi de caracteres | |
| 24 | 21-27 |  Ming-3.0-tts | 1187±13Intervalo de 95%: 1174 a 1200 | 2,418 | Oct 2026 | N/D | |
| 25 | 22-27 |  Soniox TTS Real-Time v2 | 1182±10Intervalo de 95%: 1172 a 1192 | 5,747 | Aug 2026 | US$ 14,2 /1 mi de caracteres | |
| 26 | 24-27 |  Gradium TTS (Sep 2026) | 1175±13Intervalo de 95%: 1162 a 1188 | 2,534 | Sept 2026 | US$ 47,8 /1 mi de caracteres | |
| 27 | 25-27 |  Speech 2.8 HD | 1174±10Intervalo de 95%: 1164 a 1184 | 5,515 | Feb 2026 | US$ 100,0 /1 mi de caracteres | |
| 28 | 28-32 |  SpaceXAI TTS | 1157±13Intervalo de 95%: 1144 a 1170 | 2,156 | Mar 2026 | US$ 15,0 /1 mi de caracteres | |
| 29 | 28-32 |  Falcon 2 | 1156±12Intervalo de 95%: 1144 a 1168 | 2,764 | Jul 2026 | US$ 10,0 /1 mi de caracteres | |
| 30 | 28-32 | Gradium TTS (Aug 2026) | 1154±12Intervalo de 95%: 1142 a 1166 | 2,773 | Aug 2026 | US$ 47,2 /1 mi de caracteres | |
| 31 | 28-32 | Speech 2.8 Turbo | 1153±10Intervalo de 95%: 1143 a 1163 | 5,266 | Feb 2026 | US$ 60,0 /1 mi de caracteres | |
| 32 | 30-34 |  Azure HD 2.5 | 1145±11Intervalo de 95%: 1134 a 1156 | 3,238 | Mar 2026 | US$ 22,0 /1 mi de caracteres | |
| 33 | 32-34 |  Async Flash v1.5 | 1141±11Intervalo de 95%: 1130 a 1152 | 2,937 | May 2026 | US$ 10,1 /1 mi de caracteres | |
| 34 | 32-36 | Step TTS 2 (Mar 2026) | 1137±11Intervalo de 95%: 1126 a 1148 | 3,025 | Feb 2026 | US$ 40,0 /1 mi de caracteres | |
| 35 | 33-37 | Async Pro v1.0 | 1129±12Intervalo de 95%: 1117 a 1141 | 2,792 | May 2026 | US$ 20,2 /1 mi de caracteres | |
| 36 | 34-38 | Speech 2.6 HD | 1127±11Intervalo de 95%: 1116 a 1138 | 4,512 | Oct 2025 | US$ 100,0 /1 mi de caracteres | |
| 37 | 35-38 | Speech 2.6 Turbo | 1122±10Intervalo de 95%: 1112 a 1132 | 4,760 | Oct 2025 | US$ 60,0 /1 mi de caracteres | |
| 38 | 36-42 | Simba 3.0 | 1117±11Intervalo de 95%: 1106 a 1128 | 3,237 | Feb 2026 | US$ 6,6 /1 mi de caracteres | |
| 39 | 37-43 |  Chatterbox HD | 1110±12Intervalo de 95%: 1098 a 1122 | 2,835 | — | May 2025 | US$ 40,0 /1 mi de caracteres |
| 40 | 38-43 | Fish Audio S2 Pro | 1109±11Intervalo de 95%: 1098 a 1120 | 2,874 | Mar 2026 | US$ 15,0 /1 mi de caracteres | |
| 41 | 38-43 | Speech-02-HD | 1108±11Intervalo de 95%: 1097 a 1119 | 4,817 | Mar 2025 | US$ 100,0 /1 mi de caracteres | |
| 42 | 38-45 |  TTS-1 HD | 1106±11Intervalo de 95%: 1095 a 1117 | 3,949 | Nov 2023 | US$ 30,0 /1 mi de caracteres | |
| 43 | 39-49 | Step Audio EditX (Mar 2026) | 1099±12Intervalo de 95%: 1087 a 1111 | 2,498 | Feb 2026 | N/D | |
| 44 | 41-51 |  Speech 2.6 | 1096±12Intervalo de 95%: 1084 a 1108 | 2,679 | Jul 2026 | N/D | |
| 45 | 43-51 | Turbo v2.5 | 1095±10Intervalo de 95%: 1085 a 1105 | 7,189 | Jul 2024 | US$ 50,0 /1 mi de caracteres | |
| 46 | 43-51 | Multilingual v2 | 1093±10Intervalo de 95%: 1083 a 1103 | 7,633 | Aug 2023 | US$ 100,0 /1 mi de caracteres | |
| 47 | 43-51 | Speech-02-Turbo | 1091±11Intervalo de 95%: 1080 a 1102 | 4,142 | Mar 2025 | US$ 60,0 /1 mi de caracteres | |
| 48 | 43-51 | Voxtral TTS | 1090±12Intervalo de 95%: 1078 a 1102 | 2,604 | Mar 2026 | US$ 16,0 /1 mi de caracteres | |
| 49 | 44-51 | TTS-1 | 1089±10Intervalo de 95%: 1079 a 1099 | 6,431 | Nov 2023 | US$ 15,0 /1 mi de caracteres | |
| 50 | 44-52 | OpenAudio S1 | 1087±11Intervalo de 95%: 1076 a 1098 | 4,775 | Jun 2025 | US$ 15,0 /1 mi de caracteres | |
| 51 | 44-52 | Gemini 2.5 Flash Lite TTS | 1085±11Intervalo de 95%: 1074 a 1096 | 3,447 | Dec 2025 | US$ 9,2 /1 mi de caracteres | |
| 52 | 51-58 | Flash v2.5 | 1077±10Intervalo de 95%: 1067 a 1087 | 5,952 | Dec 2024 | US$ 50,0 /1 mi de caracteres | |
| 53 | 51-59 | Journey | 1074±12Intervalo de 95%: 1062 a 1086 | 2,442 | Dec 2023 | US$ 160,0 /1 mi de caracteres | |
| 54 | 52-62 | GPT-Realtime-2 | 1073±12Intervalo de 95%: 1061 a 1085 | 2,439 | May 2026 | US$ 191,6 /1 mi de caracteres | |
| 55 | 52-62 | Studio | 1073±11Intervalo de 95%: 1062 a 1084 | 4,710 | Feb 2023 | US$ 160,0 /1 mi de caracteres | |
| 56 | 52-62 | T2A-01-HD | 1072±11Intervalo de 95%: 1061 a 1083 | 5,179 | Jan 2025 | US$ 50,0 /1 mi de caracteres | |
| 57 | 52-64 | Gemini 2.5 Flash TTS (Dec 2025) | 1068±12Intervalo de 95%: 1056 a 1080 | 3,046 | Dec 2025 | US$ 9,2 /1 mi de caracteres | |
| 58 | 52-64 |  Magpie-Multilingual 357M (Feb 2026) | 1067±12Intervalo de 95%: 1055 a 1079 | 2,390 | Feb 2026 | N/D | |
| 59 | 54-67 |  Polly Generative | 1063±11Intervalo de 95%: 1052 a 1074 | 3,914 | May 2024 | US$ 30,0 /1 mi de caracteres | |
| 60 | 53-67 |  Maya 2 Flash | 1062±13Intervalo de 95%: 1049 a 1075 | 2,290 | Jul 2026 | N/D | |
| 61 | 54-67 | Chirp 3: HD | 1062±11Intervalo de 95%: 1051 a 1073 | 3,533 | Mar 2025 | US$ 30,0 /1 mi de caracteres | |
| 62 | 55-67 |  Kokoro 82M v1.0 | 1062±11Intervalo de 95%: 1051 a 1073 | 5,539 | Jan 2025 | US$ 0,7 /1 mi de caracteres | |
| 63 | 57-69 |  Octave 2 | 1060±12Intervalo de 95%: 1048 a 1072 | 3,296 | Oct 2025 | US$ 87,5 /1 mi de caracteres | |
| 64 | 58-70 | Sonic 3 | 1056±11Intervalo de 95%: 1045 a 1067 | 3,647 | Oct 2025 | US$ 49,0 /1 mi de caracteres | |
| 65 | 59-71 |  MiMo-V2.5-TTS | 1055±12Intervalo de 95%: 1043 a 1067 | 2,219 | - | N/D | |
| 66 | 59-72 |  Coda | 1054±12Intervalo de 95%: 1042 a 1066 | 2,574 | May 2026 | US$ 50,0 /1 mi de caracteres | |
| 67 | 59-73 |  Flux TTS | 1052±13Intervalo de 95%: 1039 a 1065 | 2,337 | Aug 2026 | US$ 45,0 /1 mi de caracteres | |
| 68 | 64-74 | Async Flash v1.0 | 1048±11Intervalo de 95%: 1037 a 1059 | 4,158 | Jul 2025 | US$ 10,1 /1 mi de caracteres | |
| 69 | 63-75 | Maya 2 Global | 1048±13Intervalo de 95%: 1035 a 1061 | 2,099 | Jul 2026 | N/D | |
| 70 | 64-75 | Maya1 | 1046±12Intervalo de 95%: 1034 a 1058 | 3,412 | Jun 2025 | N/D | |
| 71 | 59-80 | OpenAudio S1 Mini | 1044±20Intervalo de 95%: 1024 a 1064 | 1,682 | Jun 2025 | US$ 15,0 /1 mi de caracteres | |
| 72 | 66-77 | Gemini 2.5 Pro (Dec 2025) | 1042±12Intervalo de 95%: 1030 a 1054 | 2,540 | Dec 2025 | US$ 18,3 /1 mi de caracteres | |
| 73 | 68-77 | Octave TTS | 1040±11Intervalo de 95%: 1029 a 1051 | 4,784 | Feb 2025 | US$ 87,5 /1 mi de caracteres | |
| 74 | 68-79 |  Higgs Audio V3 TTS | 1039±13Intervalo de 95%: 1026 a 1052 | 2,302 | Jun 2026 | US$ 15,0 /1 mi de caracteres | |
| 75 | 70-80 |  Bland Speech v3 | 1035±12Intervalo de 95%: 1023 a 1047 | 2,650 | Aug 2026 | US$ 40,0 /1 mi de caracteres | |
| 76 | 67-81 | Azure Neural | 1033±20Intervalo de 95%: 1013 a 1053 | 1,156 | Sept 2018 | US$ 15,0 /1 mi de caracteres | |
| 77 | 72-81 | MAI-Voice-1 | 1030±12Intervalo de 95%: 1018 a 1042 | 2,416 | Apr 2026 | N/D | |
| 78 | 74-81 | T2A-01-Turbo | 1028±11Intervalo de 95%: 1017 a 1039 | 5,063 | Jan 2025 | US$ 30,0 /1 mi de caracteres | |
| 79 | 73-81 | Polly Long-Form | 1027±13Intervalo de 95%: 1014 a 1040 | 2,028 | Nov 2023 | US$ 100,0 /1 mi de caracteres | |
| 80 | 76-81 | Chatterbox  | 1024±11Intervalo de 95%: 1013 a 1035 | 4,779 | May 2025 | US$ 25,0 /1 mi de caracteres | |
| 81 | 76-81 | Lightning v3.1 | 1022±12Intervalo de 95%: 1010 a 1034 | 2,426 | Mar 2026 | US$ 25,0 /1 mi de caracteres | |
| 82 | 82-83 |  Raon SpeechLM | 1007±13Intervalo de 95%: 994 a 1020 | 2,304 | Apr 2026 | N/D | |
| 83 | 83 |  Zonos-v0.1 | 1000ÂncoraFixado em 1000 como ponto de referência da escala Elo deste ranking. | 4,916 | Feb 2025 | US$ 20,0 /1 mi de caracteres | |
| 84 | 84 | Murf Speech Gen 2 | 986±12Intervalo de 95%: 974 a 998 | 5,063 | Jul 2024 | US$ 100,0 /1 mi de caracteres | |
| 85 | 85-86 |  OpenVoice v2 | 956±13Intervalo de 95%: 943 a 969 | 3,112 | — | Apr 2024 | US$ 8,3 /1 mi de caracteres |
| 86 | 85-88 | VibeVoice 1.5B | 955±14Intervalo de 95%: 941 a 969 | 2,187 | Aug 2025 | N/D | |
| 87 | 86-89 |  Neuphonic TTS | 941±14Intervalo de 95%: 927 a 955 | 2,515 | Oct 2024 | US$ 20,8 /1 mi de caracteres | |
| 88 | 86-89 | Qwen3 TTS Flash | 941±14Intervalo de 95%: 927 a 955 | 2,233 | Sept 2025 | US$ 11,0 /1 mi de caracteres | |
| 89 | 87-91 | Qwen3 TTS | 929±13Intervalo de 95%: 916 a 942 | 3,175 | Jan 2026 | US$ 13,8 /1 mi de caracteres | |
| 90 | 89-92 |  Kugel 3 | 925±13Intervalo de 95%: 912 a 938 | 4,477 | Jun 2026 | N/D | |
| 91 | 89-92 | WaveNet | 919±12Intervalo de 95%: 907 a 931 | 6,420 | Sept 2016 | US$ 4,0 /1 mi de caracteres | |
| 92 | 89-93 |  XTTS v2 | 915±15Intervalo de 95%: 900 a 930 | 2,735 | — | Nov 2023 | US$ 40,4 /1 mi de caracteres |
| 93 | 92-96 | Mist V2 | 903±15Intervalo de 95%: 888 a 918 | 3,079 | Mar 2025 | US$ 31,1 /1 mi de caracteres | |
| 94 | 93-97 | Polly Neural | 895±14Intervalo de 95%: 881 a 909 | 2,995 | Jul 2019 | US$ 16,0 /1 mi de caracteres | |
| 95 | 93-97 | Neural2 | 894±12Intervalo de 95%: 882 a 906 | 6,704 | Jun 2022 | US$ 16,0 /1 mi de caracteres | |
| 96 | 93-97 |  StyleTTS 2 | 891±15Intervalo de 95%: 876 a 906 | 2,707 | — | Jun 2023 | US$ 2,8 /1 mi de caracteres |
| 97 | 94-97 | Standard | 888±12Intervalo de 95%: 876 a 900 | 6,787 | Aug 2018 | US$ 4,0 /1 mi de caracteres | |
| 98 | 98 |  Noiz TTS | 869±15Intervalo de 95%: 854 a 884 | 2,466 | Nov 2025 | N/D | |
| 99 | 99-100 |  MetaVoice v1 | 847±18Intervalo de 95%: 829 a 865 | 1,672 | — | Feb 2024 | N/D |
| 100 | 99-100 | Polly Standard | 832±15Intervalo de 95%: 817 a 847 | 3,608 | Nov 2016 | US$ 4,0 /1 mi de caracteres |
1 O preço da API reflete o custo de gerar 1M de caracteres na API do criador do modelo com as configurações padrão do modelo
2 Pesos abertos inclui modelos que podem ser auto-hospedados usando pesos de modelo para download ou um ambiente de inferência fornecido pelo provedor. Pesos abertos para uso comercial inclui modelos cuja licença publicada permite o uso comercial, inclusive sujeito a condições da licença (por exemplo, uma assinatura do NVIDIA AI Enterprise), e exclui modelos que exigem um acordo comercial negociado separadamente.
Perguntas frequentes
Os melhores modelos de texto para fala por classificação Elo são: 1. Eleven v4 Turbo (Elo 1336), 2. Eleven v4 (Elo 1334), 3. Qwen-Audio-3.1-TTS-Plus (Elo 1305), 4. Sonic 3.6 (Elo 1296) e 5. Gemini 3.8 Flash TTS (Elo 1283). Os rankings se baseiam em votos cegos de usuários na Speech Arena.
Kokoro 82M v1.0 é o mais acessível a $0.65 por 1M de caracteres, com uma pontuação Elo de 1062. Outras opções acessíveis incluem StyleTTS 2 a $2.82 por 1M de caracteres.
Os 5 melhores modelos de texto para fala de pesos abertos são: 1. Fish Audio Drama 3 (Elo 1278), 2. Fish Audio Drama 3 Preview (Elo 1240), 3. Breeze TTS 2 (Elo 1228), 4. Ming-3.0-tts (Elo 1187) e 5. Fish Audio S2 Pro (Elo 1109).