Observed arrival · 2026-09-28
Voice Agent Bench’s Vapi–Retell phone test
A recorded comparison of two voice-agent platforms, using a dental receptionist scenario and reporting call outcomes, latency, and raw data.
○Open
⊠Login
$Paid
†Ads
✦Pretty
●Pro
◎Niche
◉Human
⚑Risk
ƒJS
- For
- Teams comparing voice-agent platforms for phone workflows
- Worth noticing
- The page links call recordings and raw JSON; 50 calls were scored, with 10 excluded for the authors’ misconfiguration.
Field notes
The test uses a dental-practice receptionist with deliberately constrained probes, including a booking tool pointed at a dead URL and a 28-second silent caller. Its latency comparison separates short requests from turns after a caller reads a phone number, rather than relying on one overall speed figure. The page links audio for scored calls and raw JSON; it also explains that 10 reached-agent calls were excluded because of the authors’ misconfiguration.
Observed signals
Read the marks
Editorial observations of this landing page, not a rating.
○OpenPublic substance visible
✦PrettyNotable craft visible
●ProPolished or operationally mature
◎NicheUnusually specific use
One card from the complete issue