US English preferences

Preference test between Eleven Multilingual V2 and Gemini 2.5 Flash Preview TTS. For each model, we use male/female voices.

Overall

E (Model B)

-0.02

G (Model A)

By Tag

E (Model B)

G (Model A)

Answers

2. G (Model A) is strongly preferred

-1

-2. E (Model B) is strongly preferred

194 (27.71%)

75 (10.71%)

140 (20.00%)

105 (15.00%)

186 (26.57%)

Tags

All tags

E (Model B)

G (Model A)

194 (27.71%)

Deep analysis

2. G (Model A) is strongly preferred

-1

-2. E (Model B) is strongly preferred

		Tags	Statistics
00:00 --:-- google/speech_us_0.wav 00:00 --:-- elevenlabs/speech_us_0.mp3	G E	address numbers narrative Man address numbers narrative Man	0.57 1.68 0.69
00:00 --:-- google/speech_us_1.wav 00:00 --:-- elevenlabs/speech_us_1.mp3	G E	address numbers narrative Woman address numbers narrative Woman	0.43 1.50 0.61
00:00 --:-- google/speech_us_2.wav 00:00 --:-- elevenlabs/speech_us_2.mp3	G E	address numbers narrative Man address numbers narrative Man	1.14 1.35 0.55
00:00 --:-- google/speech_us_3.wav 00:00 --:-- elevenlabs/speech_us_3.mp3	G E	address numbers narrative Woman address numbers narrative Woman	-1.29 1.38 0.57
00:00 --:-- google/speech_us_4.wav 00:00 --:-- elevenlabs/speech_us_4.mp3	G E	address numbers narrative Man address numbers narrative Man	0.86 1.35 0.55
00:00 --:-- google/speech_us_5.wav 00:00 --:-- elevenlabs/speech_us_5.mp3	G E	address numbers narrative Woman address numbers narrative Woman	1.29 1.38 0.57
00:00 --:-- google/speech_us_6.wav 00:00 --:-- elevenlabs/speech_us_6.mp3	G E	address numbers narrative Man address numbers narrative Man	1.29 1.38 0.57
00:00 --:-- google/speech_us_7.wav 00:00 --:-- elevenlabs/speech_us_7.mp3	G E	address numbers narrative Woman address numbers narrative Woman	-0.57 1.50 0.61

Overall

E (Model B)

-0.02

G (Model A)

By Tag

E (Model B)

G (Model A)

Answers

2. G (Model A) is strongly preferred

-1

-2. E (Model B) is strongly preferred

194 (27.71%)

75 (10.71%)

140 (20.00%)

105 (15.00%)

186 (26.57%)

Tags

All tags

E (Model B)

G (Model A)

194 (27.71%)

Deep analysis

2. G (Model A) is strongly preferred

-1

-2. E (Model B) is strongly preferred

		Tags	Statistics
00:00 --:-- google/speech_us_0.wav 00:00 --:-- elevenlabs/speech_us_0.mp3	G E	address numbers narrative Man address numbers narrative Man	0.57 1.68 0.69
00:00 --:-- google/speech_us_1.wav 00:00 --:-- elevenlabs/speech_us_1.mp3	G E	address numbers narrative Woman address numbers narrative Woman	0.43 1.50 0.61
00:00 --:-- google/speech_us_2.wav 00:00 --:-- elevenlabs/speech_us_2.mp3	G E	address numbers narrative Man address numbers narrative Man	1.14 1.35 0.55
00:00 --:-- google/speech_us_3.wav 00:00 --:-- elevenlabs/speech_us_3.mp3	G E	address numbers narrative Woman address numbers narrative Woman	-1.29 1.38 0.57
00:00 --:-- google/speech_us_4.wav 00:00 --:-- elevenlabs/speech_us_4.mp3	G E	address numbers narrative Man address numbers narrative Man	0.86 1.35 0.55
00:00 --:-- google/speech_us_5.wav 00:00 --:-- elevenlabs/speech_us_5.mp3	G E	address numbers narrative Woman address numbers narrative Woman	1.29 1.38 0.57
00:00 --:-- google/speech_us_6.wav 00:00 --:-- elevenlabs/speech_us_6.mp3	G E	address numbers narrative Man address numbers narrative Man	1.29 1.38 0.57
00:00 --:-- google/speech_us_7.wav 00:00 --:-- elevenlabs/speech_us_7.mp3	G E	address numbers narrative Woman address numbers narrative Woman	-0.57 1.50 0.61