Aller au contenu

Soniox TTS RT v1 Soniox · text-to-speech (TTS) API

Visible delay

260 ms median and 320 ms P95 time to first audio on Coval (30 days to October 22, 2026): Visible delay speed tier.

Benchmark summary: Visible delay speed tier on Coval's P95 (the voice becomes a visible part of the delay), a tie within 10 ms with Soniox TTS Real-Time v2 on slow turns.

Soniox TTS RT v1 is a proprietary text-to-speech (TTS) API from Soniox, measured by two independent benchmarks: Coval and Speko.

Speed tier

Visible delay

P95 of 320 ms or more. The voice becomes a visible part of the delay.

Median time to first audioCoval, 30 days
260 ms
P95 time to first audio (slow turns)Coval, 30 days
320 ms
Naturalness EloSpeko
1569
Listener Elo (95% interval)Artificial Analysis
not rated
Numbers and dates (0 to 1)Speko
0.35
List price per 1M charactersall sources
$13 to $14.1
  • Tagged multilingual by Coval

Data read October 22, 2026. No combined score: methodology

Gradium TTS vs Soniox TTS RT v1 Fastest TTS APIs, by speed tier

Speed

On Coval's 30-day window ending October 22, 2026 (2,962 runs from US East), Soniox TTS RT v1 starts speaking after 259.9 ms on a typical turn (median) and 320.4 ms on a slow turn (P95, the slowest one in twenty). That P95 is 14th lowest of the 32 models Coval measures, which puts it in the Visible delay tier: the voice becomes a visible part of the delay. On the median it is a tie (gap under 10 ms) with Deepgram Aura 2, Rime Mist v3 and Soniox TTS Real-Time v2.

Voice quality

Speko rates its naturalness at 1,569 Elo from blind A/B votes on phone-agent lines (13th highest of the 43 models Speko rates).

Numbers and dates

Speko's pronunciation test on numbers, dates and currency scores it 0.35 on a 0-to-1 scale (12th highest of the 30 models Speko scores). This is the accuracy measure used here; Coval's word error rate separates models too little to compare them.

Price

Coval lists $14.1 per 1M characters (since Sep 22, 2026). Speko lists about $13. List prices, as each source prints them: a range of $13 to $14.1.

Languages tested

Tested in Spanish, German and French by Artificial Analysis or Speko. A language missing here was not tested, which does not mean it is unsupported.

Benchmark measurements

Latency and accuracy — Coval (30-day window)

Mean TTFA
258.8 ms
P50 / median time-to-first-audio (TTFA)
259.9 ms
P90 TTFA
310.3 ms
P95 TTFA
320.4 ms
Latency consistency (standard deviation)
42.2 ms
Leading silence
75.5 ms
Time to first byte (network arrival)
183.3 ms
Accuracy: word error rate (WER)
2.8%
WER basis
pooled
Latency runs
2,962 runs
Accuracy runs
2,961 runs
Window
30d
Coval snapshot
2026-10-22T14:00:19.256741Z
Coval model ID
tts-rt-v1
Coval model page
benchmarks.coval.ai
Tagged multilingual
✓
Feature tags
multilingual, voice-cloning
List price per 1M characters
14.1 USD
Price listed since
Sep 22, 2026

Speko

P50 TTFA, US East
362 ms
Naturalness (Elo)
1,569
Head-to-head win rate
63%
Pronunciation robustness (0 to 1)
0.35
Voice drift (lower is steadier)
24
Price per 1M characters
13 USD
Measured on
2026-07-03
Speko model
tts-rt-v1
Speko model page
benchmarks.speko.ai

Price (all sources)

Lowest list price per 1M characters
13 USD
Highest list price per 1M characters
14.1 USD

Not collected

Similar TTS models