Cekura Bench publishes verifiable benchmarks for voice AI models. Its speech-to-speech benchmark evaluates 9 realtime voice models—including GPT Realtime 2.1, Gemini Live, Grok, and Phonic—as complete phone agents on live calls, covering 82 scenarios with three runs each. Models are ranked on reliability, data accuracy, stalled calls, response time, and cost, and every call transcript is public. The platform also covers voice agent and speech-to-text benchmarks, with text-to-speech benchmarks planned.
Cekura Bench
Speech-to-speech model benchmarks on live phone calls
Cekura Bench Introduction
Key Features
- Speech-to-speech benchmark testing 9 realtime voice models as full phone agents
- 82 scenarios run three times each on live phone calls
- Rankings across reliability, data accuracy, stalled calls, response time, and cost
- Public transcripts for every benchmark call
- Additional voice agent and STT benchmarks, with TTS benchmarks coming soon
Use Cases
- Compare speech-to-speech models before selecting one for a realtime voice agent
- Track model reliability and latency across realistic call scenarios
- Review public call transcripts to verify benchmark results
Why Startups Use It
Startups building voice agents can use Cekura Bench to compare realtime speech-to-speech models on objective, publicly verifiable criteria—reliability, accuracy, latency, and cost—before committing to a provider for their phone-based AI products.
Frequently Asked Questions
What is Cekura Bench?
Cekura Bench is a platform that publishes verifiable benchmarks for voice AI, including speech-to-speech, voice agent, and speech-to-text benchmarks, with text-to-speech benchmarks coming soon.
Which models are tested in the speech-to-speech benchmark?
The benchmark tests 9 realtime voice models, including GPT Realtime 2.1, Gemini Live, Grok, and Phonic, evaluated as complete phone agents on live calls.
How are models ranked?
Models are ranked on reliability, data accuracy, stalled calls, response time, and cost across 82 scenarios, with three runs per scenario.
Can I verify the benchmark results?
Yes, every call transcript from the benchmarks is publicly available.
More About Cekura Bench
PricingPaid
Listed DateOct 09, 2026
Authority Badge
Add our badge to your website to showcase product credibility and listing status.
Featured List