534 ms median and 917 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).
Google Chirp 3: HD is a proprietary text-to-speech (TTS) API from Google, measured by two independent benchmarks: Artificial Analysis and Coval.
Speed tier
Visible delayP95 of 320 ms or more. The voice becomes a visible part of the delay.
Data read October 22, 2026. No combined score: methodology
Gradium TTS vs Google Chirp 3: HD Fastest TTS APIs, by speed tierOn Coval's 30-day window ending October 22, 2026 (3,376 runs from US East), Google Chirp 3: HD starts speaking after 534.3 ms on a typical turn (median) and 917.2 ms on a slow turn (P95, the slowest one in twenty). That P95 is 29th lowest of the 32 models Coval measures, which puts it in the Visible delay tier: the voice becomes a visible part of the delay.
In the Artificial Analysis Speech Arena it holds 1,061 ±12 Elo (95% interval), 60th highest of the 98 models Artificial Analysis rates; scores whose intervals overlap are a tie. In its Customer Service category, Artificial Analysis places it 50th, at 1,109 Elo.
Speko has not scored its pronunciation of numbers and dates.
Coval lists $30 per 1M characters (since Aug 10, 2026). Artificial Analysis lists $30. List prices, as each source prints them.
Tested in Japanese, Hindi, Tamil, Telugu and Thai by Artificial Analysis or Speko. A language missing here was not tested, which does not mean it is unsupported.
Deepgram Aura 2 En is a proprietary text-to-speech (TTS) API from Deepgram, served by Cloudflare, measured by one independent benchmark, Coval.
Alibaba Qwen3 TTS Flash Realtime is a proprietary text-to-speech (TTS) API from Alibaba (Qwen), measured by one independent benchmark, Coval.
Smallest.ai Lightning V3.1 Pro (Jul 2026) is a proprietary text-to-speech (TTS) API from Smallest.ai, measured by three independent benchmarks: Artificial Analysis, Coval and Speko.