Jul 6, 202671 blind votes22 ranked models
TTS Leaderboard
Live Elo rankings from blind pairwise voice comparisons.
Rank
Model
1
Lightning v3.1
Smallest
1,099
±129 95% CI
100%
+50% vs even
7
7W · 0L
2
MAI-Voice-2
Microsoft/Azure
1,082
±121 95% CI
88%
+38% vs even
8
7W · 1L
3
Grok TTS
xAI
1,068
±91 95% CI
71%
+21% vs even
14
10W · 4L
4
Google Gemini 2.5 Flash TTS
Google
1,039
±129 95% CI
71%
+21% vs even
7
5W · 2L
5
MiniMax Speech-2.5 Turbo
MiniMax
1,016
±341 95% CI
100%
+50% vs even
1
1W · 0L
6
Deepgram Aura v2
Deepgram
1,015
±341 95% CI
100%
+50% vs even
1
1W · 0L
7
Murf AI Gen2
Murf
1,014
±129 95% CI
57%
+7% vs even
7
4W · 3L
8
Inworld TTS-1.5 Max
Inworld
1,005
±129 95% CI
57%
+7% vs even
7
4W · 3L
9
Maya1
Maya
1,004
±121 95% CI
50%
+0% vs even
8
4W · 4L
10
Kokoro TTS
Kokoro
1,002
±140 95% CI
50%
+0% vs even
6
3W · 3L
11
Google Gemini 3.1 Flash TTS
Google
1,000
±341 95% CI
0%
-50% vs even
0
0W · 0L
12
Hume Octave
Hume
999
±140 95% CI
50%
+0% vs even
6
3W · 3L
13
ElevenLabs Eleven v3
ElevenLabs
996
±121 95% CI
50%
+0% vs even
8
4W · 4L
14
Voxtral Mini TTS (2603)
Mistral
985
±341 95% CI
0%
-50% vs even
1
0W · 1L
Cartesia Sonic 3.5
Cartesia
977
±99 95% CI
42%
-8% vs even
12
5W · 7L
16
Amazon Polly
Amazon
975
±140 95% CI
33%
-17% vs even
6
2W · 4L
17
Google Gemini 2.5 Pro TTS
Google
973
±241 95% CI
0%
-50% vs even
2
0W · 2L
18
MiniMax Speech-02 HD
MiniMax
969
±241 95% CI
0%
-50% vs even
2
0W · 2L
19
OpenAI GPT-4o Mini TTS
OpenAI
965
±86 95% CI
38%
-12% vs even
16
6W · 10L
20
Chatterbox
Chatterbox
964
±129 95% CI
29%
-21% vs even
7
2W · 5L
21
Lightning v3.1 Pro
Smallest
946
±129 95% CI
29%
-21% vs even
7
2W · 5L
22
Qwen3 TTS Flash
Alibaba/Qwen
908
±114 95% CI
11%
-39% vs even
9
1W · 8L