Leaderboard

The leaderboard is still early. Rankings will become more reliable as more tests are completed.

Models Ranked

277

Total Matches

Active Languages

Leaderboard

Language

Provider

Min Matches

Rank	Model	Provider	ELO	Matches	Win Rate	Last Updated
1	V3	Elevenlabs	1617	34	67.6%	9/6/2026
2	Sonic Turbo	Cartesia	1591	45	55.6%	9/5/2026
3	GPT-4o-Mini-TTS	Open AI	1566	34	58.8%	5 days ago
4	Flash 2.5Provisional	Elevenlabs	1532	26	53.8%	9/6/2026
5	Sonic 3	Cartesia	1527	46	56.5%	9/19/2026
6	Gen2	Murf AI	1517	51	54.9%	10/1/2026
7	eleven_multilingual_v2Provisional	Elevenlabs	1515	1	100.0%	2/16/2026
8	Turbo 2.5Provisional	Elevenlabs	1504	26	46.2%	2/23/2026
9	Inworld TTS 1.5 Max	InWorld	1493	52	53.8%	5 days ago
10	TTS-1Provisional	Open AI	1486	23	47.8%	9/5/2026
11	TTS-1-HD	Open AI	1465	40	47.5%	10/8/2026
12	Falcon	Murf AI	1453	30	33.3%	5 days ago
13	Aura 2	Deepgram	1451	69	46.4%	9/19/2026
14	BulbulProvisional	Sarvam	1423	17	35.3%	10/1/2026
15	Inworld TTS 1.5 Mini	InWorld	1414	49	40.8%	10/1/2026
16	Amazon Polly NeuralProvisional	Amazon Polly	1364	11	18.2%	5 days ago

Not Yet Participated

Models that have not completed any blind tests yet.

Model	Provider
Speech 2.8 HD	Minimax
Speech 2.8 Turbo	Minimax

Pairwise Win Rate Matrix

Head-to-head win % (row model preferred over column model). n = number of comparisons.

Model	V3	Sonic Turbo	GPT-4o-Mini-TTS	Flash 2.5	Sonic 3	Gen2	eleven_multilingual_v2	Turbo 2.5	Inworld TTS 1.5 Max	TTS-1	TTS-1-HD	Falcon	Aura 2	Bulbul	Inworld TTS 1.5 Mini	Amazon Polly Neural
V3(Elevenlabs)	—	0.0%n=2	50.0%n=2	50.0%n=2	66.7%n=3	66.7%n=3	—	100.0%n=3	66.7%n=3	100.0%n=2	—	100.0%n=1	100.0%n=1	100.0%n=1	100.0%n=1	—
Sonic Turbo(Cartesia)	100.0%n=2	—	100.0%n=3	100.0%n=1	50.0%n=2	100.0%n=3	—	0.0%n=1	57.1%n=7	50.0%n=2	100.0%n=5	100.0%n=1	0.0%n=1	—	33.3%n=3	100.0%n=1
GPT-4o-Mini-TTS(Open AI)	50.0%n=2	0.0%n=3	—	100.0%n=1	42.9%n=7	50.0%n=4	—	—	75.0%n=4	100.0%n=1	75.0%n=4	—	33.3%n=3	—	100.0%n=4	—
Flash 2.5(Elevenlabs)	50.0%n=2	0.0%n=1	0.0%n=1	—	33.3%n=3	33.3%n=3	—	100.0%n=1	66.7%n=3	—	—	100.0%n=3	100.0%n=2	0.0%n=1	50.0%n=2	—
Sonic 3(Cartesia)	33.3%n=3	50.0%n=2	57.1%n=7	66.7%n=3	—	—	—	0.0%n=1	40.0%n=5	50.0%n=4	66.7%n=3	33.3%n=3	100.0%n=1	—	100.0%n=4	100.0%n=1
Gen2(Murf AI)	33.3%n=3	0.0%n=3	50.0%n=4	66.7%n=3	—	—	—	50.0%n=2	50.0%n=2	0.0%n=1	50.0%n=4	75.0%n=4	66.7%n=3	85.7%n=7	50.0%n=4	0.0%n=1
eleven_multilingual_v2(Elevenlabs)	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—
Turbo 2.5(Elevenlabs)	0.0%n=3	100.0%n=1	—	0.0%n=1	100.0%n=1	50.0%n=2	—	—	100.0%n=1	0.0%n=1	0.0%n=1	100.0%n=1	0.0%n=2	100.0%n=1	100.0%n=3	—
Inworld TTS 1.5 Max(InWorld)	33.3%n=3	42.9%n=7	25.0%n=4	33.3%n=3	60.0%n=5	50.0%n=2	—	0.0%n=1	—	100.0%n=3	50.0%n=6	100.0%n=2	80.0%n=5	—	16.7%n=6	100.0%n=1
TTS-1(Open AI)	0.0%n=2	50.0%n=2	0.0%n=1	—	50.0%n=4	100.0%n=1	—	100.0%n=1	0.0%n=3	—	—	100.0%n=2	33.3%n=3	—	50.0%n=2	100.0%n=1
TTS-1-HD(Open AI)	—	0.0%n=5	25.0%n=4	—	33.3%n=3	50.0%n=4	—	100.0%n=1	50.0%n=6	—	—	50.0%n=2	62.5%n=8	—	66.7%n=6	100.0%n=1
Falcon(Murf AI)	0.0%n=1	0.0%n=1	—	0.0%n=3	66.7%n=3	25.0%n=4	—	0.0%n=1	0.0%n=2	0.0%n=2	50.0%n=2	—	0.0%n=1	60.0%n=5	50.0%n=2	100.0%n=2
Aura 2(Deepgram)	0.0%n=1	100.0%n=1	66.7%n=3	0.0%n=2	0.0%n=1	33.3%n=3	—	100.0%n=2	20.0%n=5	66.7%n=3	37.5%n=8	100.0%n=1	—	—	44.4%n=9	—
Bulbul(Sarvam)	0.0%n=1	—	—	100.0%n=1	—	14.3%n=7	—	0.0%n=1	—	—	—	40.0%n=5	—	—	—	100.0%n=2
Inworld TTS 1.5 Mini(InWorld)	0.0%n=1	66.7%n=3	0.0%n=4	50.0%n=2	0.0%n=4	50.0%n=4	—	0.0%n=3	83.3%n=6	50.0%n=2	33.3%n=6	50.0%n=2	55.6%n=9	—	—	0.0%n=1
Amazon Polly Neural(Amazon Polly)	—	0.0%n=1	—	—	0.0%n=1	100.0%n=1	—	—	0.0%n=1	0.0%n=1	0.0%n=1	0.0%n=2	—	0.0%n=2	100.0%n=1	—