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 | Pricing 1 | |
|---|---|---|---|---|---|---|---|
| 1 | 1 |  Fish Audio Drama 3  | 1274±1795% interval 1257 to 1291 | 1,713 | Oct 2026 | $15.00 /1M chars | |
| 2 | 2-3 | Fish Audio Drama 3 Preview | 1239±1395% interval 1226 to 1252 | 2,747 | Oct 2026 | $15.00 /1M chars | |
| 3 | 2-3 |  Breeze TTS 2 | 1229±1395% interval 1216 to 1242 | 2,874 | Aug 2026 | $34.00 /1M chars | |
| 4 | 4 |  Ming-3.0-tts | 1185±1395% interval 1172 to 1198 | 2,587 | Oct 2026 | N/A | |
| 5 | 5-7 |  Phonon 31M (English) | 1113±1495% interval 1099 to 1127 | 1,908 | Oct 2026 | $0.00 /1M chars | |
| 6 | 5-7 | Fish Audio S2 Pro | 1106±1195% interval 1095 to 1117 | 2,977 | Mar 2026 | $15.00 /1M chars | |
| 7 | 6-8 |  Step Audio EditX (Mar 2026) | 1099±1295% interval 1087 to 1111 | 2,585 | Feb 2026 | N/A | |
| 8 | 7-8 |  Voxtral TTS | 1091±1295% interval 1079 to 1103 | 2,719 | Mar 2026 | $16.00 /1M chars | |
| 9 | 9-10 |  Magpie-Multilingual 357M (Feb 2026) | 1062±1295% interval 1050 to 1074 | 2,497 | Feb 2026 | N/A | |
| 10 | 9-10 |  Kokoro 82M v1.0 | 1061±1195% interval 1050 to 1072 | 5,618 | Jan 2025 | $0.65 /1M chars | |
| 11 | 11-13 |  Maya1 | 1046±1295% interval 1034 to 1058 | 3,496 | Jun 2025 | N/A | |
| 12 | 9-13 | OpenAudio S1 Mini | 1042±2095% interval 1022 to 1062 | 1,678 | Jun 2025 | $15.00 /1M chars | |
| 13 | 11-13 |  Higgs Audio V3 TTS | 1037±1295% interval 1025 to 1049 | 2,386 | Jun 2026 | $15.00 /1M chars | |
| 14 | 14 |  Chatterbox  | 1021±1195% interval 1010 to 1032 | 4,838 | May 2025 | $25.00 /1M chars | |
| 15 | 15 |  Zonos-v0.1 | 1000AnchorFixed at 1000 as the reference point for this board's Elo scale. | 4,979 | Feb 2025 | $20.00 /1M chars | |
| 16 | 16-17 |  OpenVoice v2 | 956±1395% interval 943 to 969 | 3,152 | — | Apr 2024 | $8.33 /1M chars |
| 17 | 16-17 |  VibeVoice 1.5B | 954±1495% interval 940 to 968 | 2,223 | Aug 2025 | N/A | |
| 18 | 18 |  XTTS v2 | 915±1595% interval 900 to 930 | 2,753 | — | Nov 2023 | $40.44 /1M chars |
| 19 | 19 |  StyleTTS 2 | 888±1595% interval 873 to 903 | 2,744 | — | Jun 2023 | $2.82 /1M chars |
| 20 | 20 |  MetaVoice v1 | 846±1895% interval 828 to 864 | 1,705 | — | Feb 2024 | N/A |
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.