Compare AI voice models
13 text-to-speech models ranked by how natural people say they sound, against what they cost to read 1,000 characters aloud. Pick a few and hear them read the same text.
Show
 Gemini 3.8 Flash TTSGoogle | 1278±15 | #5 | ≈$0.019 | ≈$0.017 | 144 | Sep 2026 | |
| Gemini 3.8 Flash-Lite TTSGoogle | 1248±14 | #8 | ≈$0.013 | ≈$0.012 | 192 | Sep 2026 | |
FAFish Audio Drama 3Fish AudioOpen weights | 1244±15 | #10 | $0.01 | $0.009 | 244 | Oct 2026 | |
FAFish Audio S2.1 ProFish Audio | 1190±14 | #18 | $0.015 | $0.013 | 127 | Jul 2026 | |
SPGrok TTSSpaceXAIZDR | 1146±14 | #27 | $0.015 | $0.013 | 97 | Mar 2026 | |
FAFish Audio S2 ProFish Audio | 1114±12 | #35 | $0.015 | $0.013 | 76 | Mar 2026 | |
 TTS-1 HDOpen AI | 1102±11 | #38 | $0.03 | $0.027 | 34 | Nov 2023 | |
| TTS-1Open AI | 1090±11 | #44 | $0.015 | $0.013 | 60 | Nov 2023 | |
FAFish Audio S1Fish Audio | 1083±11 | #47 | $0.015 | $0.013 | 55 | Oct 2025 | |
MAMAI-Voice-2-FlashMicrosoft AIZDR | — | — | $0.015 | $0.013 | — | Jul 2026 | |
MAMAI-Voice-2.1-FlashMicrosoft AIZDR | — | — | $0.015 | $0.013 | — | Oct 2026 | |
MAMAI-Voice-2Microsoft AIZDR | — | — | $0.022 | $0.02 | — | Jun 2026 | |
MAMAI-Voice-2.1Microsoft AIZDR | — | — | $0.022 | $0.02 | — | Oct 2026 |
Quality: Elo from the Artificial Analysis Text to Speech Arena, where people vote between two clips of the same text. Prices from the Vercel AI Gateway per 1,000 characters; ≈ is an estimate from audio-token prices.
Gemini and some others follow it; others ignore it
0 / 500≈ $0.00 for 3 models
Try