Voice AI BenchmarksBy

DATASETTTSCOVAL PROMPT MANIFESTACTIVE

manifest

Text prompts text-to-speech benchmark dataset

30 short text prompts that every text-to-speech model synthesizes.

Items
30
fixed public inputs
Models measured
30
last 30 days
Current leader · TTFA#1 / 30
116ms
Palabra TTS v1via Palabra

How models rank on Text prompts

Full TTS dashboard
Time to First Audio on Text promptsMilliseconds · lower is better · 30-day average on this datasetEvery TTS model measured on the Text prompts dataset, ranked on Time to First Audio.
  1. #1Palabra TTS v1116 ms
  2. #2vui124 ms
  3. #3TTS Flash 2128 ms
  4. #4TTS 2176 ms
  5. #5Blizzard235 ms
  6. #6Neural236 ms
  7. #7Mist v3256 ms
  8. #8TTS Rt v2262 ms
  9. #9Sonic 3.5274 ms
  10. #10TTS RT v1275 ms
  11. #11Dragon HD Latest310 ms
  12. #12Coda313 ms
Show all 30 models
  1. #13Aura 2328 ms
  2. #14S2.1 Pro374 ms
  3. #15Default384 ms
  4. #16Speech 2.8 Turbo411 ms
  5. #17Eleven v3 Conversational412 ms
  6. #18Grok TTS420 ms
  7. #19S1434 ms
  8. #20Flash v2.5455 ms
  9. #21Speech 2.8 HD460 ms
  10. #22Sonic 3.6466 ms
  11. #23Simba 3.2484 ms
  12. #24Chirp 3 HD512 ms
  13. #25Simba 3.0526 ms
  14. #26Falcon 2549 ms
  15. #27Lightning v3.1 Pro586 ms
  16. #28Qwen3 TTS Flash Realtime645 ms
  17. #29S2.1 Pro Free806 ms
  18. #30GPT-4o mini TTS1075 ms
median of all models · 397 ms
TTS models on the Text prompts dataset over the last 30 days, ranked on Time to First Audio.
#ModelHostTTFAWERSamples
1Palabra TTS v1Palabra116 ms5.9%14,396
2vuiFluxions124 ms8,624
3TTS Flash 2Inworld AI128 ms8,719
4TTS 2Inworld AI176 ms4.8%14,529
5BlizzardLmnt235 ms7.4%14,117
6NeuralAzure236 ms4.3%3,350
7Mist v3Rime256 ms6.3%14,524
8TTS Rt v2Soniox262 ms8,849
9Sonic 3.5Cartesia274 ms6.1%14,513
10TTS RT v1Soniox275 ms3.7%14,528
11Dragon HD LatestAzure310 ms5.1%3,350
12CodaRime313 ms5.2%14,523
13Aura 2Deepgram328 ms5.3%14,501
14S2.1 ProFish Audio374 ms10,222
15DefaultGradium384 ms4.6%14,002
16Speech 2.8 TurboMiniMax411 ms4.8%1,997
17Eleven v3 ConversationalElevenLabs412 ms6,750
18Grok TTSxAI420 ms4.7%14,524
19S1Fish Audio434 ms4.9%10,236
20Flash v2.5ElevenLabs455 ms6.7%9,259
21Speech 2.8 HDMiniMax460 ms1,995
22Sonic 3.6Cartesia466 ms560
23Simba 3.2Speechify484 ms4.7%14,525
24Chirp 3 HDGoogle512 ms5.2%14,530
25Simba 3.0Speechify526 ms5.3%14,526
26Falcon 2Murf549 ms10,229
27Lightning v3.1 ProSmallest586 ms4.4%14,515
28Qwen3 TTS Flash RealtimeAlibaba645 ms8.8%14,530
29S2.1 Pro FreeFish Audio806 ms4.7%14,464
30GPT-4o mini TTSOpenAI1075 ms4.8%14,517
Under-sampled models are excluded; tied models share a place. Dotted WER values split into substitutions, deletions and insertions on hover or tap.

Inside the dataset

30 fixed public inputs — every model is tested on exactly these.

Every prompt in the Text prompts dataset, verbatim from the manifest.
A1Your order #ORD-24589 shipped via UPS tracking 1Z999AA1234567890 and will arrive between 2:30 PM and 4:45 PM tomorrow.
A2There's a slight delay with your $347.89 order, but we expect it to ship by Friday afternoon.
A3Order status update: your items are now in "Quality Check" and should ship within 24-48 hours.
A4Hi Ms. Garcia, your appointment with Dr. Peterson is scheduled for Tuesday, March 5th at 10:30 AM.
A5I can reschedule you to either 11:15 AM on Wednesday or, let me check, 2:45 PM on Thursday.
A6Please arrive 10 minutes early and bring your insurance card and a photo ID.
A7Let's fix your Wi-Fi: unplug your router for exactly 30 seconds, then wait for the green light.
A8Your device is downloading update version 12.4.1, about 250 MB remaining, roughly 8 minutes left.
A9I'm seeing error code E-1047 here, which usually means, actually, let me guide you through the solution.
A10For security, please confirm the last 4 digits of the card ending in 8429 used for your recent purchase.
A11I need to verify your account: what's the ZIP code we have on file? Please take your time.
A12We detected 2 unusual login attempts from Chicago, Illinois yesterday around 11:30 PM.
A13Hi David, your premium subscription renews in 5 days at $29.99, would you like to continue or explore other options?
A14Based on your purchase history, you might like our flash sale: 30% off electronics through Sunday.
A15I noticed you haven't logged in since January 15th, is there anything we can help you with?
A16Your invoice shows a balance of $187.50, including the monthly fee plus tax.
A17The $24.99 charge is for the premium support package you activated on January 22nd.
A18Your auto-payment of $67.50 failed due to an expired card, would you like to update your payment method?
A19Thanks for your interest in our software, companies your size typically see 20-25% efficiency gains.
A20For a 50-person team, I'd recommend our Business plan at $149 per month rather than Basic.
A21Great timing! We're offering 20% off annual plans, that saves you $359.88 for the first year.
A22On a scale of 1 to 10, how would you rate your current pain level?
A23Have you taken any medications today? Please include prescription drugs and over-the-counter items.
A24Your symptoms sound manageable, but I'd recommend seeing Dr. Martinez within 48 hours to be safe.
A25Your PTO request for December 20th through 24th has been approved, that's 3 business days.
A26Password reset complete: your temporary password was sent to your email. Please change it within 24 hours.
A27IT ticket about your printer issue has been assigned to technician Lisa Chen.
A28Your wire transfer of $1,247.50 to account ending in 5691 was processed at 2:15 PM.
A29Hi Mr. O'Connor, I'm calling about your consultation, are you available for 20 minutes between 9 AM and noon?
A30System alert: your server usage hit 87% at 3:42 AM, but it's back to normal levels now.

Source: Coval TTS test cases · License: Apache-2.0

What it tests

Every TTS model synthesizes the same 30 prompts. Time to First Audio measures startup latency on identical text, while transcription-based WER checks whether the generated speech preserves the prompt's words.

Metrics reported for this dataset

Same datasets, prompts and metric definitions for every model, measured by Coval’s open-source runner. Full methodology on the overview.

Evaluate your own voice agent

Use Coval to test your production configuration, prompts and calls—not only the public benchmark endpoints.

Book a Demo