GradiumBENCHMARKS

Data read on November 19, 2026

Gradium TTS latency

Time to first audio is the delay between sending text and hearing the first sound. For a voice agent, it is the part of the turn the caller notices.

LatencyGradium TTSRank (median)
Median time to first audio64 ms5 / 32
P90161 ms
P95 time to first audio (slow turns)172 ms11 / 32
Leading silence (median)0 ms2 / 32
Word error rate (mean)1.7%16 / 32
Samples96

Source : Coval · Read on November 19, 2026 · gradium-tts-beta-202609

Coval TTS leaderboard, ordered by median time to first audio

Every model Coval measures, as published on its public board. Gradium is highlighted.

#ModelMedianP95Leading silenceWER
1 Fluxions vui 36 ms65 ms8 ms2.1%
2 Deepdub dd-etts-3.3 56 ms86 ms2 ms1.6%
3 Inworld inworld-tts-2-flash 60 ms118 ms0 ms1.7%
4 Nari Labs qwen3-tts-fast 63 ms101 ms10 ms1.9%
5 Gradium gradium-tts-beta-202609 64 ms172 ms0 ms1.7%
6 Palabra palabra-tts-v1 65 ms135 ms0 ms1.7%
7 Speechify simba-3.5 93 ms143 ms11 ms2.7%
8 Rime coda 99 ms153 ms24 ms1.8%
9 Speechify simba-3.0 102 ms151 ms10 ms2.2%
10 Speechify simba-3.2 103 ms149 ms10 ms2.0%
11 Baseten qwen3-tts-1.7b 118 ms143 ms37 ms1.8%
12 Murf falcon-2 141 ms235 ms31 ms1.6%
13 Inworld inworld-tts-2 160 ms213 ms57 ms1.5%
14 ElevenLabs eleven_flash_v2_5 201 ms546 ms41 ms1.9%
15 ElevenLabs eleven_v4_turbo 203 ms4291 ms64 ms1.7%
16 Deepgram flux-haley-en 205 ms449 ms53 ms1.9%
17 Assemblyai assemblyai-tts-1 209 ms328 ms121 ms1.9%
18 Airy airy-tts-v1 216 ms585 ms5 ms1.6%
19 Guava daytona-tts 261 ms297 ms3 ms1.5%
20 Xai grok-tts 262 ms454 ms50 ms1.7%
21 Deepgram aura-2-thalia-en 262 ms570 ms126 ms1.9%
22 Smallest AI lightning_v3.1_pro 284 ms531 ms56 ms1.6%
23 Cartesia sonic-3.6 292 ms398 ms138 ms1.6%
24 Soniox tts-rt-v2 301 ms353 ms81 ms3.0%
25 Soniox tts-rt-v1 305 ms363 ms75 ms2.9%
26 Fish Audio s2.1-pro 354 ms502 ms4 ms1.6%
27 Fish Audio drama-3-preview-us-east 367 ms485 ms71 ms1.4%
28 Rime mistv3 402 ms517 ms135 ms2.3%
29 Cloudflare aura-2-en 496 ms684 ms250 ms2.1%
30 Google chirp-3-hd 532 ms883 ms394 ms1.7%
31 Alibaba qwen3-tts-flash-realtime 719 ms763 ms91 ms2.7%
32 Fish Audio s2.1-pro-free 819 ms1100 ms5 ms1.6%

Official Gradium page

Coval leaderboard reading cited on gradium.ai/api and llms.txt

47 ms · P50 time to first audio (7-day window, read September 24, 2026)

read September 24, 2026

Text-to-Speech product page figures

49 ms · TTFA P50 (Speko TTS benchmark)

50 ms · TTFA P50 (Perceived TTFA, Coval benchmark)

5 ms · Latency spread (P75 − P25) (tightest of the 5 models tested, over 480 runs)

81.0% · Hard-case pass rate (500 sentences, 5 languages)

read 2026-11-19