Blog
-
Rime Prism and Phonic join our voice benchmark
First results for Rime Prism and Phonic on ten cross-customer voice scenarios, with a new Dialt Fast run and uncertainty intervals for every model.
-
The real-time voice game show
Five voice-agent systems take a quick-fire quiz. Compare correct, wrong and passed answers over one adjusted minute, with three runs each.
-
Dialt now offers on-prem voice assistants
Self-host Dialt on premises or in your own cloud, on a single GPU or across multiple GPUs, with dialects, custom voices and conversational behaviors scoped to your application.
-
Eleven Realtime joins our voice benchmark
Our first results for Mini and Max on ten cross-customer voice scenarios, with a review of tool use and response times.
-
Voice Isolation for Voice Agents
Background speaker suppression for competing speech from television, radio and announcements, with optional agent control during a call.
-
Language Switching with Voice Agents
Two ways to switch languages mid-call: keep one multilingual voice, or hand off to a voice and prompt built for the caller's language.
-
Interactive Voice Design
You can now design a virtual assistant by talking to it and adjust how it speaks, what it says, and its character in real time.
-
More Consistent Response Times for Dialt Fast and Smart
Lower p90 response times provide a more consistent caller experience while preserving benchmark-leading accuracy.
-
How to Build Evals for your Voice Agent
Build an eval suite for your own voice agent with one prompt to a coding agent, then compare Dialt Fast and GPT Live on it.
-
3 Tips for Better Voice Agent Evals
Test in text mode first, use a strong caller, and use a strong judge.
-
Dialt Preview: Fast and Smart Real-Time Voice Models
Dialt Fast and Dialt Smart against Gemini 3.8 Live and GPT Live-1 on our voice benchmark.