Fish Audio S2.1 Pro Free Fish Audio · text-to-speech (TTS) API

Name: Fish Audio S2.1 Pro Free
Author: Fish Audio

836 ms median and 1,159 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).

Fish Audio S2.1 Pro Free is a proprietary text-to-speech (TTS) API from Fish Audio, measured by one independent benchmark, Coval.

Visit website

Proprietary Coval Fish Audio Visible delay

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: 836 ms
P95 time to first audio (slow turns)Coval, 30 days: 1,159 ms
Naturalness EloSpeko: not rated
Listener Elo (95% interval)Artificial Analysis: not rated
Numbers and dates (0 to 1)Speko: not rated
List price per 1M charactersall sources: not published

Tagged multilingual by Coval

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

Gradium TTS vs Fish Audio S2.1 Pro Free Fastest TTS APIs, by speed tier

Speed

On Coval's 30-day window ending October 22, 2026 (3,366 runs from US East), Fish Audio S2.1 Pro Free starts speaking after 835.5 ms on a typical turn (median) and 1,158.8 ms on a slow turn (P95, the slowest one in twenty). That P95 is 31st lowest of the 32 models Coval measures, which puts it in the Visible delay tier: the voice becomes a visible part of the delay.

Voice quality

Neither Speko nor Artificial Analysis rates its voice quality.

Numbers and dates

Speko has not scored its pronunciation of numbers and dates.

Price

No list price is published by Coval, Speko or Artificial Analysis for this model.

Benchmark measurements

Latency and accuracy — Coval (30-day window)

Mean TTFA: 854.3 ms
P50 / median time-to-first-audio (TTFA): 835.5 ms
P90 TTFA: 1,044.1 ms
P95 TTFA: 1,158.8 ms
Latency consistency (standard deviation): 256.3 ms
Leading silence: 16.7 ms
Time to first byte (network arrival): 837.6 ms
Accuracy: word error rate (WER): 1.8%
WER basis: pooled
Latency runs: 3,366 runs
Accuracy runs: 3,364 runs
Window: 30d
Coval snapshot: 2026-10-22T14:00:19.256741Z
Coval model ID: s2.1-pro-free
Coval model page: benchmarks.coval.ai
Tagged multilingual: ✓
Feature tags: emotion-control, multilingual, voice-cloning

Not collected

Naturalness (Speko naturalness) — not measured by Speko (reading of 2026-10-22)
Listener rating (Artificial Analysis Speech Arena Elo) — not measured by Artificial Analysis (reading of 2026-10-22)
Numbers and dates (Speko pronunciation score) — not measured by Speko (reading of 2026-10-22)
Price (Artificial Analysis price per 1M characters) — not measured by Artificial Analysis (reading of 2026-10-22)

Similar TTS models

Deepgram Aura 2 En

Visible delay Proprietary · Deepgram

1027.1 ms P95

519.8 ms median

Deepgram Aura 2 En is a proprietary text-to-speech (TTS) API from Deepgram, served by Cloudflare, measured by one independent benchmark, Coval.

Coval

Google Chirp 3: HD

Visible delay Proprietary · Google

917.2 ms P95

534.3 ms median

Google Chirp 3: HD is a proprietary text-to-speech (TTS) API from Google, measured by two independent benchmarks: Artificial Analysis and Coval.

Japanese Hindi Tamil Telugu +7

Alibaba Qwen3 TTS Flash Realtime

Visible delay Proprietary · Alibaba (Qwen)

789.2 ms P95

727.7 ms median

Alibaba Qwen3 TTS Flash Realtime is a proprietary text-to-speech (TTS) API from Alibaba (Qwen), measured by one independent benchmark, Coval.

Coval