Create

Sign in to ReadmeX

or

How we handle your data: Privacy Policy

ReadmeX

Your AI reading companion

What would you like to catch up on?

Pick a briefing, or ask your own question.

Your community and people briefings stay personal to you.

Story

Artificial Analysis launches text-to-speech latency benchmark

AI summary

Artificial Analysis announced Time to First Audio, a benchmark that measures when text-to-speech models begin producing audible speech rather than when they return their first byte. It tests models daily with 12 prompts across four input lengths and reports a rolling seven-day mean; Simba 3.5 recorded 116 ms, Realtime TTS-2 186 ms, and Eleven v4 Turbo 258 ms.

In-depth insight

The benchmark provides a standardized comparison of the delay users experience before hearing a voice agent, complementing Artificial Analysis's existing quality, pronunciation and generation-speed evaluations.

Announcing Artificial Analysis Time to First Audio, measuring how quickly Text to Speech models start speaking. Among the leading models the wait ranges from 76 ms to nearly a second

Our Text to Speech benchmarks already cover quality, pronunciation and generation speed, but not how long a user waits before hearing anything — the number that decides whether a voice agent feels responsive. We measure every model daily on the same 12 prompts across four input lengths, timing audible speech rather than the first byte, and report a rolling 7-day mean.

Key results:

➤ Fastest at the frontier: @SpeechifyAI's Simba 3.5 starts in 116 ms — the highest Arena Elo of any model that begins speaking in under 200 ms. @InworldAI's Realtime TTS-2 follows at 186 ms.

➤ Turbo variants earn their name: @ElevenLabs' Eleven v4 Turbo starts in 258 ms against 653 ms for Eleven v4, at the same Arena Elo. @FishAudio's Fish Audio Drama 3 matches it at 258 ms and @cartesia's Sonic 3.6 is at 287 ms.

➤ Streaming is the dividing line: The 65 streaming models typically begin speaking within a few hundred milliseconds, a median of 376 ms. The 20 non-streaming models must finish generating before they return anything: a median of 2.0 seconds.

See more details below ⬇️

#Artificial Analysis#Speechify AI#Simba 3.5#Inworld AI#Realtime TTS-2#ElevenLabs#Eleven v4 Turbo#Eleven v4#Fish Audio#Fish Audio Drama 3#Cartesia#Sonic 3.6

Original · 1 min readRead the original
Group chat

No comments yet. Start the conversation.