What is the VoiceDuel voice arena?
VoiceDuel is a blind AI voice arena for realtime
speech-to-speech models. You write a character brief — a hiring manager, an angry
customer, a medieval knight — and two different voice models play it. You
hold a live spoken conversation with each one, then say which performance
was better. Neither model is named until after you vote.
Which AI voice models does the arena compare?
Contenders are drawn from the current generation of realtime speech-to-speech
APIs, including OpenAI gpt-realtime, xAI Grok
Voice and Alibaba Qwen Omni Realtime. These are
true speech-to-speech models — audio in, audio out from a single model —
not a text-to-speech and speech-recognition pipeline stitched together.
What does the speech arena actually rank?
Most leaderboards rank vendors. VoiceDuel ranks configurations:
the same model with different turn-detection settings, voices and
instructions. How quickly an agent decides you have finished speaking, how
it handles being interrupted, and how well it stays in character matter
more to how a voice agent feels than which company trained it — and almost
nobody measures them in public.
How is the voice model leaderboard calculated?
Votes are fitted with a Bradley-Terry model rather than Elo,
because the full vote history is available and does not need an online
approximation. Each score carries a 95% confidence interval from a
parametric bootstrap, so you can see when a ranking is still noise.
Matchmaking favours pairs that have met least often, and slot order is
counterbalanced to cancel out the tendency to prefer whoever spoke last.
Is my audio recorded?
No. Your microphone is active only during a conversation. Audio is streamed
to the voice providers being compared for the duration of the call and is
not written to disk. What is stored is the vote itself, which two
configurations were paired, and how long each conversation lasted.